It’s a nice closed loop now so it’s inevitable that ai generated content will win google search. One can pump out content, see how it scores, and iterate until they have an ai trained to top the google search results. It’s already so frustrating looking for real info. Soon it will be impossible.
That sentence isn't the longest valid English sentence, for example, this sentence is longer. In fact, as that Wikipedia article says, you can extend that sentence as well:
> Thomas Tymoczko has pointed out that there is nothing special about eight "buffalos"; any sentence consisting solely of the word "buffalo" repeated any number of times is grammatically correct.
I've tried Kagi, Brave and DDG for extended periods and ultimately their results are even worse in most cases and leave you going "hmm, I wonder what Google would give me".
It's the enshitification of the entire internet, not just Google unfortunately.
I use them all the time for work. !mdn for MDN, !dnab for .NET, !npm for NPM, !py3 for Python3 docs, !debman for Debian Manpages, !w for Wikipedia, !a for Amazon, !g for Google when you really need it.
I'm not affiliated with DDG; I just really, really love it.
Bangs are no not "asking the search engine to return all page of a particular website" but rather "directly use the search engine of that particular website".
It's a bit using the a particular search engine from your browser.
This really feels like someone living in an echo chamber. Asking my extended family if they've ever used DDG or even heard of it would yield very low numbers and I have a large extended family. Most users don't even know you can use things within Google search like "site:" to attempt to control the results.
Just look at that list of bangs you provided. It's a techie/dev's wet dream of sources. Not one of them is something somebody's mom is going to give one iota of a concern about.
Anecdotal but I found that I come across much less seo spam in general since switching to Kagi. I do use the feature to adjust certain domains’ ranking so these might just be the fruits of that.
No, but these alternative search engines are less likely to surface blogspam sites that are the biggest users of AI writing. On Kagi, you can block those domains when they come up. On Millionshort you can scope your results to exclude the top 100, top1000 or top10k domains. That clears out a lot of the garbage.
Because Google is still "good enough" for super-duper majority of people doing their searches. Very anecdotal, but sometimes I look at Google Trends / top google searches, and most people just search for very generic things (people, events, big news and etc.), which will return Wikipedia, Google Maps, Youtube, APNews, or some other massive website, and they'll be satisfied when they reach their content.
For every non-satisfactory "recipe" or "review" search, I can see people just forgetting about how they couldn't find what they wanted and since the "wins" of casual searches still overshadows "losses" of specific searches.
I'm not sure if this tracks, but outside of tech industry people, and friends who just Google everything as "my search + reddit", I haven't heard many complaints.
I despise these AI generated SEO blogs ruining the internet.
And I know people who “work” on this. They just ChatGPT prompt some shit, long posts, and paste it in. Sure its “relevant” but all reads like those recipe blogs and just repeats the same thing.
I use DDG/Brave search, but it may only be a matter of time.
How long until reputable outlets start publicly promising never to have AI write their articles? It would be great for me, the reader, if this caught on. It would let me know which outlets to take at least somewhat seriously, and which to disregard out of hand for not making this commitment.
I happen to know that, say, Microsoft, currently is publishing AI-generated news articles, with predictable results. [0] It's easy enough for me to disregard everything I see on Microsoft news, but the set of outlets I need to remember to ignore is only going to grow.
Somewhat related: StackOverflow has already banned AI-generated answers.
I knew Stack Overflow had gone back and forth on the issue, so I'm happy to see they finally listened to their actual contributors. https://stackoverflow.com/help/ai-policy
Hah, I ran across that site a while ago when looking for information about a Primus song. It had the most completely nonsensical interpretation of that song, too.
The linked article gives no examples unless you "sign up" to actually read the thing. I don't know who this is, why should I give them any consideration? Toss this in the garbage with all the rest.
I share the frustration about article paywalls, but this is unnecessarily inflammatory.
404media are good people with good repute in the infosec space (and it's the only journalism I pay for). They're all former Motherboard staff, with names you might recognize: Joseph Cox, Samantha Cole, Emanuel Maiberg, Jason Koebler.
So what is the author saying here? Seems to be: "journalist" reading other articles and writing their own derivative articles - good. Machine doing the same thing - bad.
The fact that so much of the news media produces no original reporting is the problem. The fact that they can easily be replaced by machines is a symptom.
You can't. "quantity over quality" is the most efficient path to maximum profit with minimum cost, thus the path capitalism will always optimize for, given no other incentives. Capitalism will capitalize, and AI looks as much like a magic money machine as any technology ever has.
The only thing that will stop this behavior is if AI generated content is proven to be less profitable than human generated content, which is a possibility if quality drops low enough, or the capabilities of AI versus cost stop making financial sense. But that has to happen at scale, and it might require a collapse of the tech industry in general, because everything, everywhere will be integrated with AI at that point.
Right. I'm a not so pure anti capitalist; I see the tendencies of the system and I don't like the results.
I strongly believe we can do better but it's complicated and hard.
I've been working in startups, equity, fund raising ... I have a large portfolio and do lots of investing so my views are conflicted and complicated - the most profitable stuff we're making isn't great.
Some sort of micropayment system that can efficiently send small amounts of money on one-time basis without a recurring subscription or fees to Stripe/PP/other. No crypto obviously, that just adds more work to convert to usable cash.
In Canada we can send money bank to bank via eTransfer, but there is a 50c flat fee for amounts <$100. There's no widget you can install on your site, you would need to ask people to login to your bank account or app and select the recipient manually. That level of friction would turn most people off.
I've been often wondering how difficult would it be to build a browser plugin that detected the original source of all this recycled content, and then linked to it.
I get so frustrated with articles that essentially summarize the content on the original source, even say that it was "first reported" by another publication, but then don't link to it!
Even worse, the chain of recycled content is sometimes like half a dozen links deep. It's idiotic, and a complete disservice to your readers.
I get that the competition for ad dollars is brutal, but wtf.
It hasn’t even really been journalists for a while now. There’s been other tools, just not as advanced as the current ones. Maybe they require someone with basic reading and editing skills to do a once over before publishing.
It’s not like LLMs are displacing any valuable jobs in this case. I want actual journalists to get paid to write novel stories, but this is really only a threat to industries that have been making a buck off of low effort republishing.
Starting to really wonder if the information space we call the web is in-trouble? Like it actually seems like it's happening, everything is getting polluted with AI dog shit?
Real reporters do things like pursue FOIA requests, interview people and link to original documents.
If the source you're looking at mostly provides commentary on other people's coverage (e.g. the Verge summarizing a paywalled NYT article), that's a good signal for detecting quantity over quality publications.
Not related to news articles, but I've noticed the rise of "AI influencers" that are ripping off people.
Their operation goes like this:
1) Steal/scrape content from human influencers (photos, videos)
2) Generated a face using some generative model, and deep fake it onto the data scraped from (1).
3) Automate the process with text and engagement, using some GPT model to generate the description / dialogue / etc.
4) Deploy on IG/TikTok/Onlyfans/etc.
5) Purchase followers.
Wait or pay for some "This influencer is not real!" article, which is then again re-generated and spread further by some AI news generator.
Real people see the content on their social media feed, and contribute to "organic growth". Clueless people start paying for the "exclusive" OF content and what not.
EDIT: Also, I've seen an explosion in AI/ML generated celebrity thirst traps on social media. Not outright porn, but FB/IG "friendly" photos - usually in the setting of some happening (academy awards, met gala, and general red carpet stuff). Thousands of comments and likes.
Over this past holiday my very ill uncle excitedly told me about Dr. Oz's new diabetes cure, endorsed by Tom Hanks, that he was considering. Infuriating.
I think it's already over. Fixing the destruction generative AI has wrought upon the Internet would be like unscrambling an egg. I'm an optimist when it comes to technology, but I also believe that being able to know what's real and what's not real is fundamental to survival as a species, and I think we've seeded our own destruction here. I'm confident buildings are already being built with formulas someone asked ChatGPT for help with, doctors and nurses are getting advice on how to diagnose and treat patients, subroutines are being copied and pasted into safety-critical software, etc. We put the lead in the gasoline, and we're going to face the consequences.
I mean, you can always just not rely on the internet for news and narratives and such.
And yes it is possible. It is possible to work in this economy, raise a family, eat well, and not need to be bombarded by fake anything, whether generative AI or human authored.
I think the atomic bomb is more credible as a seed of human destruction.
By the way does anyone know anything about any launch codes?
Also - have you seen the whereabouts of John Connor?
> I mean, you can always just not rely on the internet for news and narratives and such.
That commenter went on to list examples where one’s personal usage of the Internet would irrelevant to one’s wellbeing. It won’t affect one’s level of care received from doctors or one’s safety when crossing a bridge.
>I think the atomic bomb is more credible as a seed of human destruction.
Nuclear weapons absolutely are an existential risk.
>By the way does anyone know anything about any launch codes?
I don't have any top secret knowledge about current systems but have read about historical command and control systems and it's a terrifying slice of history. Now there is starting to be talk about integrating AI into those systems, which I find horrifying.
For the Internet as a whole, yes, but not for this narrow case of legitimate news organizations. If Facebook could have building full of moderators to review flagged content on their site, Google could easily fix this issue with a team to handle reports of low quality AI content. Not to mention, how most of these sites sprung up out of nowhere. That should already be a red flag for the algorithm to de-rank until a human could review. And the example of reworded blog spam with identical photos, posted at a later date, again, should be trivial for a team that cared to catch manually(flagged) or through an algorithm.
It's not over, unless Google doesn't care enough about Google News to fix it.
Humans are older as confident nonsense generators. Incompetence was around before we had machines that automate it. Of course some people will die due to machine recommendations, but I've yet to see an argument why this will be worse than deaths from overconfident humans.
The whole premise of our industry is scale and the difference it makes.
So why is it so hard to fathom that the scale this tech is enabling is making a qualitive difference and not only a quantitive one?
We place responsibility with the humans that employ the generators. And we've always needed curators. If we see a section of the web drowning in nonsense that just means that that part will become irrelevant. If Google or Facebook can't find a way to curate, they will be forgotten and only revered by some people as the good old times.
We can deplore it, but must not delude ourselves into thinking that the old web didn't have a nonsense problem. You had sites that would go unlinked because the creators were crazy. At first the thing that curbed influx of nonsense somewhat is barriers to entry such as cost and technical competence. Then the curation was reputation based with Google's pagerank as an example.
Now we will find new ways. And it hurts. It hurts like spam hurts email.
This argument could also be used to dismiss the whole "AI" industry. After all, humans have been able to answer questions and write essays for a long time!
Of course the problem lies exactly in the ability to automate incompetence.
With that take, the internet was what seeded our destruction. Doctors could get credible-sounding bad advice on the internet a decade ago, no AI needed. This is just another "don't trust everything you read in Wikipedia" moment, and it will pass.
I choose to be naively optimistic: the glut of reasonable-sounding bullshit will increase the practice of citing sources when making claims. Then we can go back to good-old-fashioned lying with statistics.
They'll have to if they don't want to be fooled. The problem is that people want to be fooled, not the fooling machine. If LLMs are better at educating than at fooling it'll be ok, if they are not it won't be ok.
We havent had an ability to tell whats real and whats not for a long time. Photos have been able to be doctored since well before photoshop, writing style imitated, signatures forged etc. This is a change of degree, thats all.
I for one am somewhat glad we are going to be forced out of our shell when it comes to the nature of truth in media. All it means is that people are going to have to have a higher standard for what they believe. Many wont, just as many today think infowars or Joe Rogan are telling them the truth, but we shouldnt cater our society to the lowest common denominator. Instead we should create a society where the more informed people can do their best to guide and educate the people more easily fooled to develop the skills to be more discriminating in what they believe.
Who was really reading/believing articles written by random people anyway? I would not take any random blog as true or useful unless I was familiar with the author and their work. How is the "AI injected" world any different? There will be more disinformation to fool those of us who believe something they read based on anything but the source. So what? Thats already a huge industry.
Fooling the easily fooled accounts for most social media and ad traffic already. So now they get more fooled or fooled more often? I think thats not good but also not catastrophic in the way that is being implied.
> Who was really reading/believing articles written by random people anyway? I would not take any random blog as true or useful unless I was familiar with the author and their work.
That works for you, but for everyone else who is less diligent in source evaluation, this now gets worse. Don't you agree that society suffers when a large proportion of people now get their information from fake sources?
The argument I am making is that this is already happening. It wont be great that the BS is more convincing and more pervasive. It also wont be the end of society and truth, leading us all to a deadly spiral of misinformation.
People should strive to be more critical, if that includes evaluating the source of whatever they listen to then I think thats great. This is something we should be doing anyway.
I think I disagree with your underlying assumption that as long as you, personally, are critically evaluating sources and are 'safe', it doesn't matter whether other people end up in the misinformation spiral.
I can be as critical as I want, if the people around me who vote, have influence on my environment, etc end up in misinformation, it has a huge impact on my life.
Yeah but those people are already taking whatever they read on the internet at face value. They already believe whatever conspiracy theory they read on tik tok or facebook. They already believe whatever fox news tells them too.
Advocating for critical and thoughtful consumption of media will become more important with the rise of AI and it will also lift up society as a whole. If society really takes this seriously we could make real progress and help many people who are currently vulnerable to misinformation become less so.
Like I said before its a matter of degree. Are you not already assuming this? Why not? Is being more critical and questioning not a good thing? Is being more educated not a good thing?
its absurd because its not going to happen. Its magical thinking to expect the general population to live past their daily life and understand these issues.
Its the type of magical thinking that also says "yes to fix obesity we just need 75% of the population to buy more expensive, less tasty, and more difficult to prepare food, and to gain more self-discipline, and to make more time for exercising. Its simply CICO". "The reason people are smoking is they need more discipline to resist their peers and all the advertising". When its clear that humans have a nature, and it doesn't really change. So in the same way, expecting a population with average 100 IQ to suddenly become discerning, diligent, credulous, scientifically minded, wary, etc. is just ludicruous. The average person is just reading headlines for like 15m at the store, in traffic, smoke break, etc. not doing a thesis every day. If they see a video of politician saying "Quit your job now, I'm handing out $1M to everyone who votes for me" they aren't going to do a deep dive to determine if its deep fake, they will just accept it.
I personally have seen two or three cases of myself or someone I know getting duped by fake info, about meaningless trivia, but just the result of a quick google like "is so and so dead" or "did so and so break up" and got wrong info, because I don't care enough to do the deep dive to invalidate it.
> We havent had an ability to tell whats real and whats not for a long time. Photos have been able to be doctored since well before photoshop, writing style imitated, signatures forged etc. This is a change of degree, thats all.
This kind of argument rubs me the wrong way. It's like saying a printing press is just a faster scriptorium, or a nuclear bomb is just a change in degree from a hand grenade. It's misleading because it ignores the monumental difference in society these later inventions brought about, which were not caused by the previous inventions, and never would have been. It ends up being an argument to not engage with the reality of the situation.
I agree. There's this weird contradiction in tech circles to hand-wave away legitimate concerns about scale (e.g., ubiquitous surveillance is just a policeman at scale), while simultaneously putting a lot of emphasis and value on whether a system is scalable.
The more complicated reality is that things qualitatively change at scale. We need to have honest sober discussions about the trade-offs and unknowns involved, instead of just YOLOing blindly into moving fast and breaking things.
I feel the issue is with human filters. Creating a seamless and believable image edit requires time and effort, regardless of how experienced the editor is. You were limited by what images you could source, how they'd fit together (does the lighting and perspective match?), and there was often a contextual paper trail people could follow ('I recognise that picture of X').
With these image generation models, the only things filtered by human involvement are the creation of the prompt, and whether or not they believe any generated image is believable. The required time and effort is minimal, and on top of that you can generate hundreds by the time a single person can make a convincing Photoshop edit.
---
In terms of being able to sift through generated images across the wider internet, I've no idea what can be done. But I feel a potential defence against them would be for news sources to publish the exact images they were sent online, keeping the original resolution, removing cropping, and retaining Exif data.
That's an interesting take. Essentially, we've been spoiled by living an a high SNR world, but that's changing, so we'll need more robust human systems.
Not to get too political, but covid and the 2020 US presidential election showed how bad we are at handling untrustworthy information.
> We havent had an ability to tell whats real and whats not for a long time. Photos have been able to be doctored since well before photoshop, writing style imitated, signatures forged etc. This is a change of degree, thats all.
But not in an automated way, that is difficult to detect, and that could impact every facet of your life. Like, yeah, photoshop was a thing, but now AI filters can "nudify" virtually any picture, quickly, and in an automated way. Ditto for generating ads, political propaganda, etc.
Most media has been coercive throughout my lifetime through the capture of distribution channels. If you control the distribution of where people consume their content, you can sway public opinion/sentiment.
For example, you might set a "narrative target" for your news outlet to report "Tech = Bad" this quarter.
Or you might purge all rural content from your network, like in the Rural Purge.
The internet has blown that wide open. Many people, especially younger demographics, are curating their own content now. Instead of "news networks," many of my younger friends talk about individual citizen journalists. Those journalists go by many names, vloggers, podcasters, youtubers, etc.
I've noticed that, in the past few years, the people proclaiming the loudest that "Journalism is Dead!" are doing so because they've ignored where the craft of journalism has migrated to: the internet. They look at the hollow dying husks of organizations that were built around capturing distribution channels, husks that have burned their credibility over years of blatantly coercive reporting, and somehow feel nostalgia for those husks.
The legacy distribution channels have, by-in-large, lost. The only metrics they appear to be relevant in are the ones that conveniently exclude the distribution channel they've lost to: the internet.
With the rise of OSS AI, I'm _extremely_ bullish on personalized moderation and curation.
In the next few years, I want a personal AI that automatically mines all of my social media feeds, email, text messages, calendars, subscribed channels, podcasts, etc. and gives me a personalized curated view.
I want that AI to be highly-biased, and highly-aligned. But the measure of bias and alignment is _me_. I want it to have my bias. I want it to have my alignment. I want it to be fine-tuned on me. And I want it to go churn through a lifetime of content daily and bubble up the most valuable/meaningful content for me.
I think we are moving into a period where the ability for people to be coercive by controlling media distribution channels is coming to an end. The rise of personalized AI will kill it.
And this is before I get into generative AI disrupting the larger entertainment industry for the better!
And where will these models run?
And who exactly will create the content?
Your personal AI will just consume a shitload of content that the powers with most compute generates. Reality is gone, it won't even be accessible via the shadows on the cave wall.
Locally, on my devices. Where property law says I own the fine-tuned model and the device.
> And who exactly will create the content?
Anyone and everyone? I don't understand what you mean. What you just posted here is content. You produced content. I assume you will continue doing that moving into the future.
> Your personal AI will just consume a shitload of content that the powers with most compute generates
I think that assumes a random walk through the internet? Content discovery probably shouldn't be random?
As your AI curates content, you give it ongoing feedback about the quality of the content it gives you, (feedback based on your highly biased definition of "good" and "bad" content). Your AI learns what is "good" to show you and what is "bad" to show you.
As your AI finds "identities" and "content channels" that produce reliably high-quality content that your highly-biased and highly-aligned AI can curate for you, I'd imagine your AI would prioritize searching those content spaces for quality content first.
For example, I'd like to think my personalized AI would learn to surface up interesting conversations and content from Hacker News.
As content channels get "overrun" by noise, the AI is filtering out the noise, the signal-to-noise ratio drops, and the AI slowly lowers the priority of those content channels in favor of channels that consistently produce content you find valuable to consume.
The key here is that the AI is ultimately accountable to me, overseen by me, controlled by me, fine-tuned on signals generated by me, etc.
Imagine you had an army of 1 million personal executive assistants who could scour the world looking for information relevant to you, and give you a curated feed of useful information on an ongoing basis. That sounds like a pretty cool future to me.
And if someone wants to spend their time squandering their wealth to accumulate enough hardware to run 1 trillion personal executive assistants to generate a bunch of valueless (or value-negative), coercive, uninteresting dribble on the internet? Cool. Let them squander their wealth. My executive assistants will filter that out for me.
Personally, I want to be challenged regularly. I want AI's that are biased based on the direction I need to be challenged in the moment, which I would mark as "good." I want an AI that can challenge my assumptions as I work by surfacing relevant research papers that show I might be following the wrong path, which I would mark as "good."
I want AIs that can adopt the personas of someone I'm reading that I disagree with, who has long been dead, so we can have a conversation over beers and sort out some of my personal philosophy.
And I want all of that on my terms.
To ground this desire in classical terms, I want to pick the books I read. I have a bookshelf full of books. I have books I agree with. I have books I disagree with. But I only have books that make me think. Being able to pick my own books doesn't mean I want to live in an echo chamber. It means I want to cultivate my own personal philosophy on my terms.
An ai persona of someone dead will be filled with guesswork, fakes, and inaccuracies, just like today's generative ai. The technology is literally guessing at the missing peices, with no ability to verify accuracy. Enjoy conversations with people while they live, because the dead don't talk explicitly.
The only thing that's real is the actual stuff in the physical world that you can see and hear with your own senses. Mechanical/digital/artistic reproductions (text/images/videos) were never real and I'm a bit baffled by how all those "loss of consensus truth" doomer takes overlook that fact. 1980s-2010s mass media is not the be-all end-all of possible worldviews and human communication/coordination patterns.
Google is almost useless since a few years now. Generative AI might have made it worse, and more obvious, but Google not going anything about aggressive SEO practices was already the beginning of the end.
With all their recent extra efforts in AI research (Gemini etc) I wonder if they’ve given up on search being the core of their being and are committing to AI products ?
It might. But the key actor here is the person who used an LLM to paraphrase another article. The fact that Google indexed it is hardly even relevant. The claim that Google is "boosting" it is meaningless.
If you read the article, you'll see that they're specifically talking about Google News including these rip-off sites, not Google merely indexing them. I would consider that a form of "boosting," especially considering that some actual news sites that create their own content are excluded.
YouTube has been running ads non-stop showing deep fakes of well-known cryptocurrency icons like Michael Sailor and Anatoly Yakovenko asking people to send some crypto to an address and get 2x back. It feels like the bottom is officially falling out.
I'm gonna say something that probably some people won't like. I'm impressed by how much money the industry is moving by this so called "AI" movement and literally having zero concrete improvement on one's every day life (and by this I mean common people, not developers/techies).
I haven't seen, nor witnessed anything that you can benefit from all this AI hype. All that has done is opening the door for polluting the web with crap. And we keep on and on with the madness of training, wasting money (and resources) on gpus, generating tb of data. For what? To generate text, images?
If someone can actually point out a real example on the real world how all this madness helps to a better world, much appreciated, because honestly I don't see it.
yes I've definitely thought quite a few times about open-sourcing our company code (AI note-taker for remote meetings). But then the hard thing is Kubernetes on top...
And as always not sure our "closed-source", "employee only" code would be useful to anyone not working with us :/
Google, Microsoft and Zoom will all have it integrated soon enough and many companies will pay extra for it because of “security and privacy” guarantees the tech companies will make.
Yes sure. But that’s not anywhere near ground breaking. Also the math here is akin to those who believe that if they didn’t smoke they’d be driving a Ferrari. In reality owning a Ferrari is way more expensive than your monthly cigarettes and comes with belles and whistles (luxury taxes, maintenance, etc).
AI could change our lives completely over the course of a decade but if each change is just a little improvement that builds on the last one, the whole time people could grumble that "that's not anywhere near ground breaking."
Using RAG on a set of docs for a project it's possible to get good answers to specific questions (and ask follow ups, etc) - really speeds up understanding, like having someone who knows their stuff answering questions for you all the time.
Making meeting notes, summarizing discussions - frees the note taker and produces a nice record of what happened.
Yeah but someday, the ChatGPT service will fail, people will send the bullet points directly because that's what they're accustomed to writing now, recipients will be pleasantly surprised to see the summarizer working automatically, and countless hours of wasted human effort will be saved with no further computation required.
I don't know about anyone else, but for me the value in taking notes is the actual process of taking the notes, not the end product of the notes themselves.
Having AI (or, indeed, a human colleague) do it for me is a bit pointless, as it eliminates 90% of the value.
I don't think making the world better is their goal, so I'm not surprised you aren't seeing those results. Their goal is to either make more money, cut costs, or both.
I believe there is more to it than just problematic or “almost correct” code examples after 7 retries with ELI5 explanations.
But part of me believes this is another “self-driving car” thing where in 10 years we’ll still be in exact same place.
Today I saw an interview from a Greek YT economist who was asked about BTC. His answer was “it’s been 10+ years that has bene around. We know by now that is not new or groundbreaking in any substantial way” and I agree. There are many tech that “if this and that” but we never seem to reach “that” point.
You can't tell people you're "24 month from full self driving in every driving condition" every 12 months for a decade an expect people to continue drinking the koolaid
It's the same for "AI", it's happening in cycle since the 60s, nobody learns a damn thing, it's borderline a pump and dump scheme at that point
No we can't. Instead we can expect people to start dismissing it, despite the fact that it keeps steadily improving, so when it finally gets good enough it takes everybody by surprise.
Not sure if this is related to a better world, but I know that some of my family members use it to write nice, "thoughtful" words to family and friends. Like for birthdays and stuff. There's something in me which strongly goes against this kind of use, but there's your example.
They also use it to learn about people and places, a bit like using Wikipedia, which they never used.
Regarding non-dev uses, I use it to simplify and "neutralize" a text I'm writing to others, like getting a second opinion on the text. For example for the homeowners association.
For example now I passed `What does this mean? "Regarding non-dev uses, I use it to simplify and "neutralize" a text I'm writing to others, like getting a second opinion on the text. For example for the homeowners association."` and read the answer to see if it coveys my message properly. The answer was helpful: "[...] This could be particularly useful in ensuring that the message is clear, diplomatic, and devoid of unintended connotations, which is important in formal or communal communications like those in a homeowners association context."
> If someone can actually point out a real example on the real world how all this madness helps to a better world, much appreciated, because honestly I don't see it.
US government killbots will bring peace in our time!
It's just taking a little longer than we had hoped to perfect the "friend or foe" module
ChatGPT is far and away the best tutor I have ever had. It starts to fail once you get past mid level concepts but its ability to provide links to other sources, explain things, not get offended when you call it out as wrong about something (which human teachers certainly do) and many other aspects make it very helpful. Specifically its useful in expanding your breadth of knowledge on a subject. Less so depth.
Regardless I have learnt things and taken on projects I never would have attempted without it.
They alluded in their past comment that human teachers make errors too which students can identify. So perhaps if students don't implicitly trust human tutors, then so too can they corroborate ChatGPT against reality and solidify their understanding.
I don't know about where you're from but in my country to teach requires several years of qualifications, assessment and continuing observation by an independent regulator
and there are consequences if you're caught teaching rubbish
not to say they are perfect, but there is some level of quality control
meanwhile ChatGPT just bullshits endlessly with no possible way to punish OpenAI for it
I think perhaps you had a bad experience with GPT at some point or didnt ask it to provide sources that you then read through. Either way it is far more reliable than you seem to believe.
I've just listened to the latest Numberphile podcast, and one of many surprising anecdotes from Donald Knuth was that as a kid he'd been disillusioned by maths when he "proved" that 1=0 and his teacher couldn't find the mistake.
> meanwhile ChatGPT just bullshits endlessly with no possible way to punish OpenAI for it
Sure there are, and more importantly there's also ways to "punish" the models themselves — that's what the "thumbs down" and "regenerate response" buttons in the chat UI are for.
Human teachers can be audited without advanced knowledge in maths, computer science, and information theory to discover what went wrong.
Human teachers can be held accountable.
A Human teacher has limited scale thus limited damage over bad information with more safeguards than "Big Corp promises not to be bad"
A human teacher can realize their own mistakes.
Human teachers have complex incentive structures that can modify their behavior.
Many humans (even ones with technical training) tend to see computers as infailible in first principals of math and carry that forward to deep learning which is where the math becomes tough to audit, couple that with LLMs affinity for overcoming the uncanny valley enough to seem like a rational human actor because of the datasets being used and you have "magic". This is before you even add financial incentive.
Not saying these issues in deep learning can't be tackled but the disengenous hype-train and the downvoting because people point out flaws with the new internet ATM is wearing thin.
There are entire industries devoted to helping humans overfit on standardized tests. Humans love to overfit, it's often easy and faster to get what you want--which in the case of education, is usually credentials, not knowledge.
> Humans love to overfit, it's often easy and faster to get what you want
The problem is Goodhart's. Tests aren't used to measure your ability to overfit them, but to measure your general knowledge. Obviously the latter is exceptionally difficult to measure so we settle for an easier metric. The problem comes from believing that the measurement well aligns with the thing we set out to measure. The same is for humans as is for machines and it is why it is insane to say something like "look, GPT did well on the BAR exam, it should be able to practice law." Because we explicitly do not want a legal system that follows the letter of the law (rather spirit of the law) and it's clear which one GPT would do. Yes, there's a problem that humans go this way too, but it's something we can solve if we are willing to admit the fuzziness of our metrics.
> hallucinations are rare for me, even at graduate-level material.
How do you know? If you're using it as a tutor you're using it to get knowledge that you don't know. Sure, it doesn't get things that are obviously (to you) wrong, but what about subtle things? At the graduate level subtly dominates and many concepts are counter intuitive or non-obvious in the first place (otherwise they wouldn't be expert knowledge).
I know because I check often with what I already know, or official material/explanations written by humans.
Gpt helps me connect dots and uncover useful details I didn't previously have, and that official explanations didn't include. And just about every time I've double-checked, it's right, even when it didn't make sense to me (counter-intuitive/non-obvious due to my mistakes in reading comprehension or previously erroneous concepts in my mind).
Okay, with this added nuance I'm not worried. But I think we should be explicit about this because it is difficult to distinguish if people are doing this or not and it is quite reasonable to believe that they are not.
I have no issues with anyone using GPT as a __noisy__ tutor. But we just must be clear it is easy to get fooled, difficult to validate, and easy to cause GPT to change its answer. There's a high rate of people using it as a non-noisy tutor, so we can't be implicit.
FWIW, I use GPT in a very similar way, and frequently. Especially with the decline of Google and especially for high level topics. It functions very well as a fuzzy search and often even when wrong (which is frequent for me when asking high level math questions, think post PDEs) there's still words or phrases that help me significantly narrow down a Google search.
Which any Google Search engineer should not be reading this as "replace Google Search with GPT/Bard"!!!!! They work in tandem!!! I CANNOT STRESS THIS ENOUGH and do not want to see the hell that that creates.
The level of cautiousness many feel about the risks of AI hallucinations feels a bit excessive in a world where many people can easily obtain chainsaws without the need to fulfill any significant prerequisites
This is a cynical take on what I think is a meaningful benefit that the above commenter is talking about. If you use ChatGPT to learn concepts and approach it with the caveat that you should make sure you understand what you are trying to learn (something that helps to verify that what is being presented to you by the system is "accurate"), it's a fantastic resource. That isn't to say it's true for learning all things, but I'm generally impressed and optimistic for ChatGPT to be as valuable a teaching/learning resource as any other I might find on the web.
I have more confidence in a teacher, especially in niche subjects. One of the big reasons why is that I can ask a teacher a question, they will infer what I am trying to ask, and not easily flip flop. Because a teacher has confidence and GPT doesn't. A teacher will even find their own mistakes without prompting (though of course, everyone needs it sometimes. Just a teacher won't say that a kilogram of feathers and a pound of bricks weigh the same and then explain how 1 kg is 2.54 lbs).
Sure, no system is perfect, but not all imperfections are equal, and I'm tired of pretending like they are.
> Sure, no system is perfect, but not all imperfections are equal, and I'm tired of pretending like they are.
I think this is the actual issue, rather than the other specifics you raise — my anecdote of human mistakes includes having a 44kHz/16-bit audio recording of a whole university lecture where the lecturer insisted that nobody can record mocap joint positions for a mere dozen joints at a mere 24Hz/8-bit for more than a few minutes because that is too much data to store, and it was their confidence that kept them from updating.
But, back to the point.
Every instance of any specific AI release (GPT or LLM or whatever else) are identical; errors are therefore necessarily correlated. Humans get an advantage from our imperfections being un-equal, we benefit from the wisdom of the crowd (with conditions) and second opinions are meaningfully independent (again with conditions), but not so for any
specific AI.
That's not true even before mixture of experts because the systems are stochastic. They are only identical if we were to always initialize with the same starting seed and rng_state.
While they do indeed have a random number generator, I think the main strengths and weaknesses of these models (and humans) are in the structures rather than at the token temperature stage — I expect a human to answer any given question slightly differently at different times, but asking the same doctor for a second opinion as the first is considered the setup for a joke rather than a sensible idea.
This is the assumption of a flat and smooth latent manifold. If either of those conditions are not met, then seeds can have a wild effect on the model output. There's only bounds on the latent structure being locally smooth or locally flat btw. This is the underpinning of such works as the Lottery Hypothesis but there's a whole rabbit hole to dig down here and it is why Alpha Go did so poorly in round 2. This is an important part of the discussion of generalizability btw. I do not think your analogy works out and it is inconsistent with your prior comment, because you're right that anyone asking GPT is akin to anyone asking the same doctor. The second doctor is a second model (or potentially a second GPT checkpoint depending on how we want to consider divergence).
Very good. Now apply that healthy skepticism to the contexts where you should have been applying it in the past but haven't been, from Stack Overflow to Wikipedia to your high school and college teachers to the evening news.
Chatgpt will not get offended when you call it out. Even when it was right and you are wrong. You call it out and now your incorrect notion is gospel as far as chatgpt is concerned. How can that possibly teach you anything?
This is another false belief that a lot of people have. ChatGPT sometimes wont correct false statements but just as often it will. This is also why we read the source and dont believe anything anyone says without proof.
Professors and chatbots are both wrong all the time and its up to the student to extract useful knowledge.
That's the hardest thing for me is that it nearly always accepts your corrections. It will often change answers even if you just ask it if it is sure or suggest it might be wrong. It's weird to say, but GPT has no confidence. Weird because it spouts things out without concern for their validity, and we might call that "confidently incorrect." But it isn't confidence because GPT won't stick to its guns like a human would. So even some of our common notions are completely divorce from how the machine actually works.
My major concern when using GPT and especially with hearing how common it is for people to use it as a tutor, is that being a novice you explicitly want a confidently correct tutor. There are certainly good aspects of GPT but you're here to gain knowledge. Being a novice explicitly means you are going to have difficulties recognizing incorrect information so you put yourself in a concerning position of becoming confidently incorrect.
You might want to specify that you're talking about generative AI specifically, as opposed to LLMs in general. I abhor stuff like what's in the OP, but don't really have a problem with using it as, for instance, a research tool, with human experts actually confirming its output: https://www.newscientist.com/article/2411374-ai-comes-up-wit...
The foundamental question is: is making people more productive "making world better"?
For example, ChatGPT has already brought huge benefits to all the non-native speakers who work for/with companies in English speaking countries. Does it count as "making world better"? Or it makes lives of white-collar workers in these countries worse, since it brought more competition in?
You are on the right track. The truth is whether its making the world better or not is unknowable. To know it you have to know the future. Which is why people who strive to make value judgement always do so by predicting the future. But even if you get it right, the future has future still. Whats immediately bad might be good for the big picture, vice versa. This is doubly true for complex things like AI.
All I'm saying is that with thing like AI its entirely dishonest to say its making the world better with a straight face. It is what it is. Don't fool yourself by thinking you're doing it for 'the greater good'.
First, I love how “AI” is just a meaningless term here. The topic was generative text and images, and now you’ve brought up image classification, a different problem. But hey, computers!
> I'm gonna say something that probably some people won't like. I'm impressed by how much money the industry is moving by this so called "AI" movement and literally having zero concrete improvement on one's every day life (and by this I mean common people, not developers/techies).
> I haven't seen, nor witnessed anything that you can benefit from all this AI hype. All that has done is opening the door for polluting the web with crap. And we keep on and on with the madness of training, wasting money (and resources) on gpus, generating tb of data. For what? To generate text, images?
> If someone can actually point out a real example on the real world how all this madness helps to a better world, much appreciated, because honestly I don't see it.
The comment I was replying to did not say LLM or Generative AI. It didn’t specify LLM and simply asked for examples where AI is making good.
Instead of saying, “is there a source?”, or taking 2 seconds to Google “4 year cancer detection with AI” people are lighting up the comments about failed AI from IBM from ten years ago.
Also re Watson. Top Google article for the search terms “Watson medical ai” leads to this article, which contains the best response to your criticism:
> I don’t think that the failure of Watson means that artificial intelligence isn’t ready to make significant improvements and changes in health care. I think it means the way that they approached it is a cautionary tale that lays out how not to do it.
Whether LLMs deserve all this hype is debatable, but whether there is merit in AI technologies being used today is not. There are some pretty good use cases in medical, pharma and biotech, agricultural, etc. Few of them are based on LLMs but that wasn’t what the commented was actually asking. They were asking if “the AI hype” had any real applications, and it does.
LLMs are just the manifestation of the Gartner hype cycle anyways. And we haven’t gotten anywhere near the terminus for them either, with almost every vendor and open source project planning to at least double model size in the next year as hardware and GPU capabilities expand. I think we have yet to see what LLMs can really do and will be much more pleased with them in 2-3 years as they mature.
> AI is able to detect cancer 4 years earlier than a doctor traditionally would.
Which "AI" ?
Which "cancer" ?
What's the false positive/negative rate ?
Regardless of all that, if it detected it 4 years earlier it means it happened at least 4 years ago and have nothing to do with the LLM hype gp's criticising.
> I haven't seen, nor witnessed anything that you can benefit from all this AI hype.
AI translators are opening up the whole world to communication. I can talk to and do business with people even if we speak no common language. I can parse information from any language now.
Incredible that you hackers are down voting this, I answered his question exactly. AI translators are very useful tools for international business and communication. Some languages were impossible for machine translating up until a few years ago. This has a real impact on the world for non-tech people. It's not a toy.
Small increase in economic productivity. Like the invention of the word processor or spreadsheet. It's hard to quantify but it trickles into the economy in the form of lower prices for the same goods (or prices that aren't as high as they otherwise would have been).
I'm not impressed that much either (yet) but this is a little unfair, ChatGPT is barely a year old and while there is a lot of hype and talk it's also having concrete effects on the real world -- it was a key sticking point for the Writer's Guild and SAG-AFTRA during their strikes last year, and why they lasted so long.
Kind of reminds me of this screenshot someone took of the fastest growing subreddits back in 2023, and all of them were AI-related, but here's the kicker, I think 8 out of 10 of them had to do with AI waifus and porn.
To be fair that's just human nature, which AI can't really change. Even if we achieve AGI, a lot of the most asked questions will probably be money/sex related.
I'm impressed by how much money the industry is moving by this so called "AI" movement and literally having zero concrete improvement on one's every day life
One small thing: Amazon.com appears to be using an LLM to provide a summary of all of a product's reviews. But aside from that, I haven't seen anything useful out of "AI" yet.
> If someone can actually point out a real example on the real world how all this madness helps to a better world, much appreciated, because honestly I don't see it.
Anecdote time.
My sister, who doesn't care much about computers, found ChatGPT useful in her job, as it gave her a template to build from that included things she'd not thought of initially. I've just listened to the latest Numberphile podcast, and was surprised to hear Donald Knuth saying similar about a handful of the responses it gave him when he was playing with it.
I play around with the GPT models regularly, even before ChatGPT came out, and it broadens my abilities by producing code in languages I don't have much professional experience with, but which I can verify.
I sometimes use it for translations, and have occasionally tried using it as a foreign language conversation partner (it's OK at that, one limitation is that it keeps going back to English, though this is also a problem I face in real life with actual humans).
The way I imagine it (sometimes spot on, sometimes hilariously wrong) large companies could eventually have really good company representatives that answer phone/email/chat immediately and are deeply familiar with the company and the customer. Many complaints can be converted into polished product/feature proposals and bug reports. Eventually they should be able to have smaller constraints, draw up plans, fill agendas and make decisions within their budget. A bit like algorithms are trusted to trade stocks.
Seems funny, have a bot from acme inc, give it a list of job descriptions, assign a budget. acme then figures out how to extract as much money as possible.
In the end, humanoid robots running the company, doing meetings and interviews all over the world simultaneously.
Walking around the factory, hey Jim, how is the wife? Oh, things are slow today, I'm going to give you the rest of the day off. Not real empathy but simply the most profitable manipulation.
What you’re noticing is a sort of active “commodity fetish” with this AI stuff.
Just like NFTs only bigger. Just like 80’s era diet pills. Or 90’s era whatevers.
It’s a facet of capitalism to drive consumption and move to profit as quickly as possible. I think we are still in the “gauge and gamble” phase, but eventually in a season or to we will see it hit and bilk the general public.
We all know “AI” has been a “thing” since ML or Soundex or PID loops, or small models, etc.
It’s just “finally” hit a downhill allowing some momentum to build.
Prompted by comments foreshadowing the end of search engines being useful, I wonder if we'll see a return to pre-search engine practices given that in the hypothetical we would come full loop from low signal to noise back to low signal to noise.
So, in the larger context of how AI will be used by people seeking a quick buck, was Sam Altman right in the need for something like World Coin, or are there other proposed alternatives?
215 comments
[ 3.8 ms ] story [ 227 ms ] threadSEO spam has been a thing for a long time that has made their way into top rank spots for some queries making it frustratingly hard to find answers.
If you make the war against SEO spam in general they'll catch themselves in the crossfire.
https://en.wikipedia.org/wiki/Buffalo_buffalo_Buffalo_buffal...
> Thomas Tymoczko has pointed out that there is nothing special about eight "buffalos"; any sentence consisting solely of the word "buffalo" repeated any number of times is grammatically correct.
Anyone who wants to go down that rabbit hole, see https://en.wikipedia.org/wiki/List_of_linguistic_example_sen...
It's the enshitification of the entire internet, not just Google unfortunately.
I use them all the time for work. !mdn for MDN, !dnab for .NET, !npm for NPM, !py3 for Python3 docs, !debman for Debian Manpages, !w for Wikipedia, !a for Amazon, !g for Google when you really need it.
I'm not affiliated with DDG; I just really, really love it.
It's a bit using the a particular search engine from your browser.
Just look at that list of bangs you provided. It's a techie/dev's wet dream of sources. Not one of them is something somebody's mom is going to give one iota of a concern about.
For every non-satisfactory "recipe" or "review" search, I can see people just forgetting about how they couldn't find what they wanted and since the "wins" of casual searches still overshadows "losses" of specific searches.
I'm not sure if this tracks, but outside of tech industry people, and friends who just Google everything as "my search + reddit", I haven't heard many complaints.
And I know people who “work” on this. They just ChatGPT prompt some shit, long posts, and paste it in. Sure its “relevant” but all reads like those recipe blogs and just repeats the same thing.
I use DDG/Brave search, but it may only be a matter of time.
https://oldtimemusic.com/the-meaning-behind-the-song-the-rat...
Fuck this new web.
I happen to know that, say, Microsoft, currently is publishing AI-generated news articles, with predictable results. [0] It's easy enough for me to disregard everything I see on Microsoft news, but the set of outlets I need to remember to ignore is only going to grow.
Somewhat related: StackOverflow has already banned AI-generated answers.
[0] https://edition.cnn.com/2023/11/02/tech/microsoft-ai-news/in...
https://en.wikipedia.org/wiki/404_Media
404media are good people with good repute in the infosec space (and it's the only journalism I pay for). They're all former Motherboard staff, with names you might recognize: Joseph Cox, Samantha Cole, Emanuel Maiberg, Jason Koebler.
The fact that so much of the news media produces no original reporting is the problem. The fact that they can easily be replaced by machines is a symptom.
It's often piecemeal work where a freelancer is writing articles and they get paid based on all the metrics that we hate.
It's set up to be as close to lorem ipsum as possible.
How would you radically transform the industry in a way that those polluters go away?
The only thing that will stop this behavior is if AI generated content is proven to be less profitable than human generated content, which is a possibility if quality drops low enough, or the capabilities of AI versus cost stop making financial sense. But that has to happen at scale, and it might require a collapse of the tech industry in general, because everything, everywhere will be integrated with AI at that point.
I strongly believe we can do better but it's complicated and hard.
I've been working in startups, equity, fund raising ... I have a large portfolio and do lots of investing so my views are conflicted and complicated - the most profitable stuff we're making isn't great.
In Canada we can send money bank to bank via eTransfer, but there is a 50c flat fee for amounts <$100. There's no widget you can install on your site, you would need to ask people to login to your bank account or app and select the recipient manually. That level of friction would turn most people off.
I'm not saying that's not a solution just because it hasn't worked, just that the sauce is more subtle and harder to get right
I get so frustrated with articles that essentially summarize the content on the original source, even say that it was "first reported" by another publication, but then don't link to it!
Even worse, the chain of recycled content is sometimes like half a dozen links deep. It's idiotic, and a complete disservice to your readers.
I get that the competition for ad dollars is brutal, but wtf.
It’s not like LLMs are displacing any valuable jobs in this case. I want actual journalists to get paid to write novel stories, but this is really only a threat to industries that have been making a buck off of low effort republishing.
My Instagram feed is under attack too.
Everything is. Even totally unimportant stuff. Sad beans.
If the source you're looking at mostly provides commentary on other people's coverage (e.g. the Verge summarizing a paywalled NYT article), that's a good signal for detecting quantity over quality publications.
Their operation goes like this:
1) Steal/scrape content from human influencers (photos, videos)
2) Generated a face using some generative model, and deep fake it onto the data scraped from (1).
3) Automate the process with text and engagement, using some GPT model to generate the description / dialogue / etc.
4) Deploy on IG/TikTok/Onlyfans/etc.
5) Purchase followers.
Wait or pay for some "This influencer is not real!" article, which is then again re-generated and spread further by some AI news generator.
Real people see the content on their social media feed, and contribute to "organic growth". Clueless people start paying for the "exclusive" OF content and what not.
EDIT: Also, I've seen an explosion in AI/ML generated celebrity thirst traps on social media. Not outright porn, but FB/IG "friendly" photos - usually in the setting of some happening (academy awards, met gala, and general red carpet stuff). Thousands of comments and likes.
I mean, you can always just not rely on the internet for news and narratives and such.
And yes it is possible. It is possible to work in this economy, raise a family, eat well, and not need to be bombarded by fake anything, whether generative AI or human authored.
I think the atomic bomb is more credible as a seed of human destruction.
By the way does anyone know anything about any launch codes?
Also - have you seen the whereabouts of John Connor?
That commenter went on to list examples where one’s personal usage of the Internet would irrelevant to one’s wellbeing. It won’t affect one’s level of care received from doctors or one’s safety when crossing a bridge.
it's not me I worry about
it's all the facebook users, who can also vote
Nuclear weapons absolutely are an existential risk.
>By the way does anyone know anything about any launch codes?
I don't have any top secret knowledge about current systems but have read about historical command and control systems and it's a terrifying slice of history. Now there is starting to be talk about integrating AI into those systems, which I find horrifying.
Ok, but what do I do about the other 159,690,456 people that vote in this country?
It's not over, unless Google doesn't care enough about Google News to fix it.
We can deplore it, but must not delude ourselves into thinking that the old web didn't have a nonsense problem. You had sites that would go unlinked because the creators were crazy. At first the thing that curbed influx of nonsense somewhat is barriers to entry such as cost and technical competence. Then the curation was reputation based with Google's pagerank as an example.
Now we will find new ways. And it hurts. It hurts like spam hurts email.
Of course the problem lies exactly in the ability to automate incompetence.
I choose to be naively optimistic: the glut of reasonable-sounding bullshit will increase the practice of citing sources when making claims. Then we can go back to good-old-fashioned lying with statistics.
This isn't hypothetical. We already have bullshit citations and bullshit articles today. The only thing remaining is to stick the two together.
I for one am somewhat glad we are going to be forced out of our shell when it comes to the nature of truth in media. All it means is that people are going to have to have a higher standard for what they believe. Many wont, just as many today think infowars or Joe Rogan are telling them the truth, but we shouldnt cater our society to the lowest common denominator. Instead we should create a society where the more informed people can do their best to guide and educate the people more easily fooled to develop the skills to be more discriminating in what they believe.
A large enough change in degree becomes a qualitative change.
If what once required nation states budget is suddenly free and in your pocket, it’s potentially a tectonic shift.
Fooling the easily fooled accounts for most social media and ad traffic already. So now they get more fooled or fooled more often? I think thats not good but also not catastrophic in the way that is being implied.
That works for you, but for everyone else who is less diligent in source evaluation, this now gets worse. Don't you agree that society suffers when a large proportion of people now get their information from fake sources?
People should strive to be more critical, if that includes evaluating the source of whatever they listen to then I think thats great. This is something we should be doing anyway.
I can be as critical as I want, if the people around me who vote, have influence on my environment, etc end up in misinformation, it has a huge impact on my life.
Advocating for critical and thoughtful consumption of media will become more important with the rise of AI and it will also lift up society as a whole. If society really takes this seriously we could make real progress and help many people who are currently vulnerable to misinformation become less so.
Honestly this an absurd take, and I like AI…
I personally have seen two or three cases of myself or someone I know getting duped by fake info, about meaningless trivia, but just the result of a quick google like "is so and so dead" or "did so and so break up" and got wrong info, because I don't care enough to do the deep dive to invalidate it.
This kind of argument rubs me the wrong way. It's like saying a printing press is just a faster scriptorium, or a nuclear bomb is just a change in degree from a hand grenade. It's misleading because it ignores the monumental difference in society these later inventions brought about, which were not caused by the previous inventions, and never would have been. It ends up being an argument to not engage with the reality of the situation.
The more complicated reality is that things qualitatively change at scale. We need to have honest sober discussions about the trade-offs and unknowns involved, instead of just YOLOing blindly into moving fast and breaking things.
With these image generation models, the only things filtered by human involvement are the creation of the prompt, and whether or not they believe any generated image is believable. The required time and effort is minimal, and on top of that you can generate hundreds by the time a single person can make a convincing Photoshop edit.
---
In terms of being able to sift through generated images across the wider internet, I've no idea what can be done. But I feel a potential defence against them would be for news sources to publish the exact images they were sent online, keeping the original resolution, removing cropping, and retaining Exif data.
Not to get too political, but covid and the 2020 US presidential election showed how bad we are at handling untrustworthy information.
Just like between a fork and a nuke, not sure it's a very solid argument
But not in an automated way, that is difficult to detect, and that could impact every facet of your life. Like, yeah, photoshop was a thing, but now AI filters can "nudify" virtually any picture, quickly, and in an automated way. Ditto for generating ads, political propaganda, etc.
Most media has been coercive throughout my lifetime through the capture of distribution channels. If you control the distribution of where people consume their content, you can sway public opinion/sentiment.
For example, you might set a "narrative target" for your news outlet to report "Tech = Bad" this quarter.
Or you might purge all rural content from your network, like in the Rural Purge.
The internet has blown that wide open. Many people, especially younger demographics, are curating their own content now. Instead of "news networks," many of my younger friends talk about individual citizen journalists. Those journalists go by many names, vloggers, podcasters, youtubers, etc.
I've noticed that, in the past few years, the people proclaiming the loudest that "Journalism is Dead!" are doing so because they've ignored where the craft of journalism has migrated to: the internet. They look at the hollow dying husks of organizations that were built around capturing distribution channels, husks that have burned their credibility over years of blatantly coercive reporting, and somehow feel nostalgia for those husks.
The legacy distribution channels have, by-in-large, lost. The only metrics they appear to be relevant in are the ones that conveniently exclude the distribution channel they've lost to: the internet.
With the rise of OSS AI, I'm _extremely_ bullish on personalized moderation and curation.
In the next few years, I want a personal AI that automatically mines all of my social media feeds, email, text messages, calendars, subscribed channels, podcasts, etc. and gives me a personalized curated view.
I want that AI to be highly-biased, and highly-aligned. But the measure of bias and alignment is _me_. I want it to have my bias. I want it to have my alignment. I want it to be fine-tuned on me. And I want it to go churn through a lifetime of content daily and bubble up the most valuable/meaningful content for me.
I think we are moving into a period where the ability for people to be coercive by controlling media distribution channels is coming to an end. The rise of personalized AI will kill it.
And this is before I get into generative AI disrupting the larger entertainment industry for the better!
Locally, on my devices. Where property law says I own the fine-tuned model and the device.
> And who exactly will create the content?
Anyone and everyone? I don't understand what you mean. What you just posted here is content. You produced content. I assume you will continue doing that moving into the future.
> Your personal AI will just consume a shitload of content that the powers with most compute generates
I think that assumes a random walk through the internet? Content discovery probably shouldn't be random?
As your AI curates content, you give it ongoing feedback about the quality of the content it gives you, (feedback based on your highly biased definition of "good" and "bad" content). Your AI learns what is "good" to show you and what is "bad" to show you.
As your AI finds "identities" and "content channels" that produce reliably high-quality content that your highly-biased and highly-aligned AI can curate for you, I'd imagine your AI would prioritize searching those content spaces for quality content first.
For example, I'd like to think my personalized AI would learn to surface up interesting conversations and content from Hacker News.
As content channels get "overrun" by noise, the AI is filtering out the noise, the signal-to-noise ratio drops, and the AI slowly lowers the priority of those content channels in favor of channels that consistently produce content you find valuable to consume.
The key here is that the AI is ultimately accountable to me, overseen by me, controlled by me, fine-tuned on signals generated by me, etc.
Imagine you had an army of 1 million personal executive assistants who could scour the world looking for information relevant to you, and give you a curated feed of useful information on an ongoing basis. That sounds like a pretty cool future to me.
And if someone wants to spend their time squandering their wealth to accumulate enough hardware to run 1 trillion personal executive assistants to generate a bunch of valueless (or value-negative), coercive, uninteresting dribble on the internet? Cool. Let them squander their wealth. My executive assistants will filter that out for me.
Personally, I want to be challenged regularly. I want AI's that are biased based on the direction I need to be challenged in the moment, which I would mark as "good." I want an AI that can challenge my assumptions as I work by surfacing relevant research papers that show I might be following the wrong path, which I would mark as "good."
I want AIs that can adopt the personas of someone I'm reading that I disagree with, who has long been dead, so we can have a conversation over beers and sort out some of my personal philosophy.
And I want all of that on my terms.
To ground this desire in classical terms, I want to pick the books I read. I have a bookshelf full of books. I have books I agree with. I have books I disagree with. But I only have books that make me think. Being able to pick my own books doesn't mean I want to live in an echo chamber. It means I want to cultivate my own personal philosophy on my terms.
I haven't seen, nor witnessed anything that you can benefit from all this AI hype. All that has done is opening the door for polluting the web with crap. And we keep on and on with the madness of training, wasting money (and resources) on gpus, generating tb of data. For what? To generate text, images?
If someone can actually point out a real example on the real world how all this madness helps to a better world, much appreciated, because honestly I don't see it.
US employees spend 18 hours / week in meetings so that is a noticeable improvement IMHO.
And as always not sure our "closed-source", "employee only" code would be useful to anyone not working with us :/
Using RAG on a set of docs for a project it's possible to get good answers to specific questions (and ask follow ups, etc) - really speeds up understanding, like having someone who knows their stuff answering questions for you all the time.
Making meeting notes, summarizing discussions - frees the note taker and produces a nice record of what happened.
Person A: ChatGPT, expand these bullet points into a ten page presentation.
(sends to person B)
Person B: ChatGPT, scan this ten page presentation and give me the main points.
Having AI (or, indeed, a human colleague) do it for me is a bit pointless, as it eliminates 90% of the value.
But part of me believes this is another “self-driving car” thing where in 10 years we’ll still be in exact same place.
Today I saw an interview from a Greek YT economist who was asked about BTC. His answer was “it’s been 10+ years that has bene around. We know by now that is not new or groundbreaking in any substantial way” and I agree. There are many tech that “if this and that” but we never seem to reach “that” point.
It's the same for "AI", it's happening in cycle since the 60s, nobody learns a damn thing, it's borderline a pump and dump scheme at that point
I wouldn't even be surprised if personal cars are banned entirely before we reach self driving cars
I want us to be still having these discussions in 2034, because that would mean I'm still capable of gainful employment for the next ten years.
But a mere 3 years ago, what ChatGPT is doing now — even with the mistakes — was SciFi level of natural language processing.
They also use it to learn about people and places, a bit like using Wikipedia, which they never used.
Regarding non-dev uses, I use it to simplify and "neutralize" a text I'm writing to others, like getting a second opinion on the text. For example for the homeowners association.
For example now I passed `What does this mean? "Regarding non-dev uses, I use it to simplify and "neutralize" a text I'm writing to others, like getting a second opinion on the text. For example for the homeowners association."` and read the answer to see if it coveys my message properly. The answer was helpful: "[...] This could be particularly useful in ensuring that the message is clear, diplomatic, and devoid of unintended connotations, which is important in formal or communal communications like those in a homeowners association context."
I would never look them in the eyes again if I found out someone did this to me.
Which makes those expressions have about as much value as the ones in greeting cards: almost none.
US government killbots will bring peace in our time!
It's just taking a little longer than we had hoped to perfect the "friend or foe" module
Regardless I have learnt things and taken on projects I never would have attempted without it.
> Regardless I have learnt things
who cares what we learn is true as long as we learnt something?
and there are consequences if you're caught teaching rubbish
not to say they are perfect, but there is some level of quality control
meanwhile ChatGPT just bullshits endlessly with no possible way to punish OpenAI for it
> meanwhile ChatGPT just bullshits endlessly with no possible way to punish OpenAI for it
Sure there are, and more importantly there's also ways to "punish" the models themselves — that's what the "thumbs down" and "regenerate response" buttons in the chat UI are for.
Human teachers can be audited without advanced knowledge in maths, computer science, and information theory to discover what went wrong.
Human teachers can be held accountable.
A Human teacher has limited scale thus limited damage over bad information with more safeguards than "Big Corp promises not to be bad"
A human teacher can realize their own mistakes.
Human teachers have complex incentive structures that can modify their behavior.
Many humans (even ones with technical training) tend to see computers as infailible in first principals of math and carry that forward to deep learning which is where the math becomes tough to audit, couple that with LLMs affinity for overcoming the uncanny valley enough to seem like a rational human actor because of the datasets being used and you have "magic". This is before you even add financial incentive.
Not saying these issues in deep learning can't be tackled but the disengenous hype-train and the downvoting because people point out flaws with the new internet ATM is wearing thin.
There's a reason GPT-4 scores so highly on advanced exams like the USMLE.
yes, it's called overfitting
The problem is Goodhart's. Tests aren't used to measure your ability to overfit them, but to measure your general knowledge. Obviously the latter is exceptionally difficult to measure so we settle for an easier metric. The problem comes from believing that the measurement well aligns with the thing we set out to measure. The same is for humans as is for machines and it is why it is insane to say something like "look, GPT did well on the BAR exam, it should be able to practice law." Because we explicitly do not want a legal system that follows the letter of the law (rather spirit of the law) and it's clear which one GPT would do. Yes, there's a problem that humans go this way too, but it's something we can solve if we are willing to admit the fuzziness of our metrics.
How do you know? If you're using it as a tutor you're using it to get knowledge that you don't know. Sure, it doesn't get things that are obviously (to you) wrong, but what about subtle things? At the graduate level subtly dominates and many concepts are counter intuitive or non-obvious in the first place (otherwise they wouldn't be expert knowledge).
Gpt helps me connect dots and uncover useful details I didn't previously have, and that official explanations didn't include. And just about every time I've double-checked, it's right, even when it didn't make sense to me (counter-intuitive/non-obvious due to my mistakes in reading comprehension or previously erroneous concepts in my mind).
I have no issues with anyone using GPT as a __noisy__ tutor. But we just must be clear it is easy to get fooled, difficult to validate, and easy to cause GPT to change its answer. There's a high rate of people using it as a non-noisy tutor, so we can't be implicit.
FWIW, I use GPT in a very similar way, and frequently. Especially with the decline of Google and especially for high level topics. It functions very well as a fuzzy search and often even when wrong (which is frequent for me when asking high level math questions, think post PDEs) there's still words or phrases that help me significantly narrow down a Google search.
Which any Google Search engineer should not be reading this as "replace Google Search with GPT/Bard"!!!!! They work in tandem!!! I CANNOT STRESS THIS ENOUGH and do not want to see the hell that that creates.
Sure, no system is perfect, but not all imperfections are equal, and I'm tired of pretending like they are.
I think this is the actual issue, rather than the other specifics you raise — my anecdote of human mistakes includes having a 44kHz/16-bit audio recording of a whole university lecture where the lecturer insisted that nobody can record mocap joint positions for a mere dozen joints at a mere 24Hz/8-bit for more than a few minutes because that is too much data to store, and it was their confidence that kept them from updating.
But, back to the point.
Every instance of any specific AI release (GPT or LLM or whatever else) are identical; errors are therefore necessarily correlated. Humans get an advantage from our imperfections being un-equal, we benefit from the wisdom of the crowd (with conditions) and second opinions are meaningfully independent (again with conditions), but not so for any specific AI.
(Possible twist: Mixture of Experts).
Professors and chatbots are both wrong all the time and its up to the student to extract useful knowledge.
My major concern when using GPT and especially with hearing how common it is for people to use it as a tutor, is that being a novice you explicitly want a confidently correct tutor. There are certainly good aspects of GPT but you're here to gain knowledge. Being a novice explicitly means you are going to have difficulties recognizing incorrect information so you put yourself in a concerning position of becoming confidently incorrect.
For example, ChatGPT has already brought huge benefits to all the non-native speakers who work for/with companies in English speaking countries. Does it count as "making world better"? Or it makes lives of white-collar workers in these countries worse, since it brought more competition in?
All I'm saying is that with thing like AI its entirely dishonest to say its making the world better with a straight face. It is what it is. Don't fool yourself by thinking you're doing it for 'the greater good'.
Early research shows that AI is able to detect cancer 4 years earlier than a doctor traditionally would.
So let’s bring it back to generative systems. IBM Watson was a complete failure, generating nonsensical diagnoses and treatments. https://slate.com/technology/2022/01/ibm-watson-health-failu...
> I haven't seen, nor witnessed anything that you can benefit from all this AI hype. All that has done is opening the door for polluting the web with crap. And we keep on and on with the madness of training, wasting money (and resources) on gpus, generating tb of data. For what? To generate text, images?
> If someone can actually point out a real example on the real world how all this madness helps to a better world, much appreciated, because honestly I don't see it.
The comment I was replying to did not say LLM or Generative AI. It didn’t specify LLM and simply asked for examples where AI is making good.
Instead of saying, “is there a source?”, or taking 2 seconds to Google “4 year cancer detection with AI” people are lighting up the comments about failed AI from IBM from ten years ago.
Here is the source:
https://www.unilad.com/technology/ai-detected-breast-cancer-...
Also re Watson. Top Google article for the search terms “Watson medical ai” leads to this article, which contains the best response to your criticism:
> I don’t think that the failure of Watson means that artificial intelligence isn’t ready to make significant improvements and changes in health care. I think it means the way that they approached it is a cautionary tale that lays out how not to do it.
https://slate.com/technology/2022/01/ibm-watson-health-failu...
Whether LLMs deserve all this hype is debatable, but whether there is merit in AI technologies being used today is not. There are some pretty good use cases in medical, pharma and biotech, agricultural, etc. Few of them are based on LLMs but that wasn’t what the commented was actually asking. They were asking if “the AI hype” had any real applications, and it does.
LLMs are just the manifestation of the Gartner hype cycle anyways. And we haven’t gotten anywhere near the terminus for them either, with almost every vendor and open source project planning to at least double model size in the next year as hardware and GPU capabilities expand. I think we have yet to see what LLMs can really do and will be much more pleased with them in 2-3 years as they mature.
https://digialps.com/the-first-ai-medical-device-that-can-de...
That’s like pointing to GM’s 1990s electric car while explaining that electric cars haven’t delivered.
Which "AI" ?
Which "cancer" ?
What's the false positive/negative rate ?
Regardless of all that, if it detected it 4 years earlier it means it happened at least 4 years ago and have nothing to do with the LLM hype gp's criticising.
> proceeds to give absolutely no example nor source
Yeah sure.
Here’s a link:
https://www.unilad.com/technology/ai-detected-breast-cancer-...
AI translators are opening up the whole world to communication. I can talk to and do business with people even if we speak no common language. I can parse information from any language now.
Incredible that you hackers are down voting this, I answered his question exactly. AI translators are very useful tools for international business and communication. Some languages were impossible for machine translating up until a few years ago. This has a real impact on the world for non-tech people. It's not a toy.
Are we absolutely sure about that?
One small thing: Amazon.com appears to be using an LLM to provide a summary of all of a product's reviews. But aside from that, I haven't seen anything useful out of "AI" yet.
Anecdote time.
My sister, who doesn't care much about computers, found ChatGPT useful in her job, as it gave her a template to build from that included things she'd not thought of initially. I've just listened to the latest Numberphile podcast, and was surprised to hear Donald Knuth saying similar about a handful of the responses it gave him when he was playing with it.
I play around with the GPT models regularly, even before ChatGPT came out, and it broadens my abilities by producing code in languages I don't have much professional experience with, but which I can verify.
I sometimes use it for translations, and have occasionally tried using it as a foreign language conversation partner (it's OK at that, one limitation is that it keeps going back to English, though this is also a problem I face in real life with actual humans).
Seems funny, have a bot from acme inc, give it a list of job descriptions, assign a budget. acme then figures out how to extract as much money as possible.
In the end, humanoid robots running the company, doing meetings and interviews all over the world simultaneously.
Walking around the factory, hey Jim, how is the wife? Oh, things are slow today, I'm going to give you the rest of the day off. Not real empathy but simply the most profitable manipulation.
There is a project that tracks/curates "AI" in the context of bad use or real world (mis-)application [1].
[1] https://incidentdatabase.ai/
Just like NFTs only bigger. Just like 80’s era diet pills. Or 90’s era whatevers.
It’s a facet of capitalism to drive consumption and move to profit as quickly as possible. I think we are still in the “gauge and gamble” phase, but eventually in a season or to we will see it hit and bilk the general public.
We all know “AI” has been a “thing” since ML or Soundex or PID loops, or small models, etc.
It’s just “finally” hit a downhill allowing some momentum to build.