Ask HN: Is anyone else getting AI fatigue?

205 points by graderjs ↗ HN
AI is great. ChatGPT is incredible. But I feel tired when I see so many new products being built that incorporate AI in some way, like "AI for this..." "AI for that..." I think it misapplies AI. But more than that, it's just too much. Right? Right? Anyone else feel like this? Everything is about ChatGPT, AI, prompts or startups we can build with that. It's like the crypto craze all over again, and I'm a little in dread of the shysters again, the waste, the opportunity cost of folks pursuing this like a mad crowd rather than being a little more thoughtful about where to go next. Not a great look for the "scene" methinks. Am I alone in this view?

378 comments

[ 3.6 ms ] story [ 287 ms ] thread
It's just the regular wantrepreneurs wave when a new shiny thing is released, we had the same with crypto, give it a few months they'll crawl back from where they came
I think Machine learning already went through the trough of disillusion around 2016-2018 in computer vision and around 2018-2020 for voice assistants.

I think we're now past that and people can see that tools like ChatGPT are powerful enough to be applied in many pre existing contexts and industries in unpredictable and inventive ways without huge amounts of manual configuration, which makes it more exciting.

Machine learning is way too generic of a term. Everything from linear regressions to neural models is technically "machine learning".

Language models are right now at the very top of the peak of inflated expectations. It's still too early to tell what the real impact will be, but it won't be even remotely close to what you read on the headlines.

Far more impressive technology (like Wolfram Alpha) has existed for almost a decade now, and it's directly comparable to language models for many applications.

My guess is they will end up being something like Rust. Very cool to look at, little impact on your day-to-day.

ML/AI is a repeat offender (for that matter, so is The Almighty Blockchain; it managed a few hype cycles under slightly different identities; blockchain, ICO, NFTs, and so on). Remember in the late 90s when Microsoft and Apple both appeared fully convinced that voice would be the primary interface with computers imminently? There was also a large brief chat agent bubble a few years back.
If you can jump around without prediction which point is next, the hype cycle is useless. There are terms ppl use for things that are en vogue. There is no hype cycle.
Yes, many people are feeling AI fatigue. AI can be overwhelming and many people feel like they are being bombarded with information that they don't understand. People are also concerned about how AI is being used and its potential implications for privacy and security.

Sorry, I couldn't help; that is the ChatGPT response to your question. More informatively, AI is clearly at the height of inflated expectations. It will provide a helpful tool. However, it will not push people out of jobs. Furthermore, right now it gives a much better search experience than Google, as it is not yet filled with ads or has been gamed extensively by SEO. It is doubtful this will stay like this in the future.

I could tell by the start of the second sentence.
ChatGPT overuse of the word "overwhelming" and a couple other similar words is very characteristic. I think it comes from the "political correctness"/"provide kind answers" prompts it is bombarded with during training
That first paragraph. It is big thing that machine can generate something like that but in reality it feel like it brings just noise. Not sure why anyone expect this to improve SEO noise.
I think it’s less like the crypto craze than the PC, web, or smartphone “crazes”, where businesses starting incorporating each of the above into everything.

In other words, if you’re fatigued already, I have some bad news regarding the rest of your life.

ChatGPT is great, but it's being hyped up so much right now. We've got AI bros coming in the scene trying to sell everybody a new product. Before the crypto craze, it was big data. I probably missed something in between.
ChatGPT has certainly made a splash, but it's part of a larger trend. I started following developments in modern AI when Kevin Kelly tweeted[1] this in 2016:

> The business plans of the next 10,000 startups are easy to forecast: Take X and add AI.

I think the AI hype cycle isn't done building. A few days ago, Paul Graham tweeted[2] this:

> One of the differences between the AI boom and previous tech booms is that AI is technically more difficult. That combined with VC funds' shift toward earlier stage investing with less analysis will mean that, for a while, money will be thrown at any AI startup.

[1]: https://twitter.com/kevin2kelly/status/718166465216512001

[2]: https://twitter.com/paulg/status/1623060319403905026

It was actually briefly ML again; there was a chat agent VC funding bubble before the main crypto VC funding bubble.
We are not there yet. And crypto is not dead, too.
* not yet
Well, eventually either the governments will survive, or crypto will. Governments (as nation states) exist longer than crypto, but they have their share of problems as of lately.
Are you seriously trying to imply that Bitcoin has a greater chance of survival than the concept of a nation state? Is this a real opinion you're holding in your head?
The concept of a nation state exists for around 300 years, and currently in a crisis. Crypto exists for 15 years (also currently in a crisis, to be fair). I think both governments and crypto will survive somehow, but not in their current forms. There might be distributed states, corporate states, fiat states, confederations etc; the “nation” aspect of the nation-state will not survive too long in my opinion (i.e. some people who are alive now will see the end of nation-states). Crypto will be transformed too, perhaps becoming more utilitarian and less reliant on competitive adversarial selection.

This is all speculative, of course, but I have seen the fall of the Soviet system, and I am well aware that forms of government are not eternal.

tl;dr but yes. Crypto of the future will look more or less similar to the crypto of today. Governments of the future will look nothing like today’s nation-states.

How is crypto not dead?
the same way stock markets arent dead.
Total Cryptocurrency Market Cap is over a trillion dollars
(comment deleted)
Well, I am old enough to remember the cryptowinters of 2015, 2018, and early 2020. Not the first, not the last. BTC/USD is still around 22,000, which looks very much not zero to me.
If crypto is dead, then I'm sure you wouldn't mind gifting me several bitcoin :D
You're too smart to be here ??? ;))
You are hardly alone, but you are also likely in the midst of an actual paradigm shift, so the "ecosystem" does what ecosystems do. ie: herds stampede, flocks flock, scavengers scavenge, parasites uhh parasite.

It will be increasingly tiresome until it becomes commonplace, then the disastrous consequences will become the next tedium.

It seems to part of a trend to prompt us and in some ways mold us into giving responses, under the probably sincere guise of convenience, like the famous Word paper clip. I think some find that useful, but it does tend to take away agency. This may be the slide to the AI's taking over :-)
We're in the middle of an AI hype. Much like with previous hypes (crypto etc), time will tell whether it was worth it. Unless you're chasing gold or selling shovels the only thing to do is just to wait it out.
I had that since I was doing my masters in data science (5 years ago?). I love the models, the statistics and just the cleverness of everything but I just can't stand the "scene" anymore and moved almost entirely away from it. It's not as exciting as it was anymore.

When I started with the topic I watched a documentary with Joseph Weizenbaum ([1]) and felt weirded out that someone would step away from such an interesting and future-shaping topic. But the older I get, the more I feel that technology is not the solution to everything and AI might actually make more problems than it solves. I still think Bostrom's paperclip maximizer ([2]) is lacking fundamental understandings of the status quo and just generated unnecessary commotion.

[1] http://www.plugandpray-film.de/en/ [2] https://www.lesswrong.com/tag/paperclip-maximizer

you mentioned you moved almost entirely from the AI / data science "scene." Where did you move to?
Being a CTO (doing manager stuff), regular coding. By moving away I also meant I don't follow along anymore and don't contribute to the projects I did so in the past. I just lost interest.
Yes, PoW crypto is now a much more concrete example of the potential damage from poorly aligned utility functions, as well as the challenges in containing a system once it is released.

I'm finding the current hype cycle very frustrating from both sides. On one side there is frequent overplaying current capabilities, and cherry picked examples given as it they're representative. On the other side there is an over simplistic "AI is evil" reaction. It's hard to deny that progress in the past few years greatly exceeds expectations and could make a significant improvement to individual creativity and learning, as well as how we cooperate but so much of the discussions are fear based.

Same here, didn’t do a masters, but worked as a data scientist for a good while.

> I love the models, the statistics and just the cleverness of everything but I just can't stand the "scene" anymore

This really sums up my feelings too.

You may want to try to mentally reframe it so it doesn't bother you as much because it's not going away any time soon.
If you're tired of AI now you're gonna hate where we are going. Strap in!

(…or take a good step back from the news cycle, check in once or twice a week instead of several times daily. News consumption reduction is good for mental health.)

This is something any crypto-bro would have told you in 2017.
Really don't understand the constant crypto comparisons. We have one technology that hasn't provided any benefits whatsoever in 10 years and one that has provided real utility from day one. One deserves the hype, the other doesn't.
Bitcoin has provided hundreds of billions in value, chatGPT has provided me with one hundred times the spam.

I'm actually optimistic about both crypto and AI, but I see the authors point. I really don't think the comparison is hard to spot between the AI hype and, say, the NFT hype from 1 year ago.

A lot of people are claiming that these technologies will imminently change everything, fundamentally. In reality, both of them are just neat things that give us a glimpse of what the future may hold, and hold a bunch of promise, but aren't really changing anything fundamentally. Not yet, at least.

I never even fully recovered from the "Facebook, but for X" fad.

At least all the previous crazes didn't threaten to replace humans, so I suppose this tech hype bubble is arguably even more irritating.

The world has some cool new toys and I'm glad people are playing with them. I hope it only becomes increasingly accessible with time. Yes, there's going to be a ton of snake oil and disappointments if you listen to the people looking to get rich quick, but I'm excited for what might come from it in the end.

In the meantime, all the attention and media is easing people into thinking about some difficult questions that we may end up having to deal with sooner than we'd like.

The hype can be annoying, and I'm sure they'll be suckers who lose a lot of money chasing it, but I'm also sure AI will get better, and be better understood too, as a result of all of the attention and attempts to shoehorn it into new roles and environments.

What winds me up is the mis-branding, sometimes deliberate sometimes not (which one is worse?!), of basic computer processing as "AI".

It's not AI it's an IF statement for crying out loud :-(

But this is the industry we're in, and buzzword-driven headlines and investment are how it goes.

Actual proper AI getting some attention makes a pleasant change tbh :-)

I feel like AI-scientists themselves are partially to blame for this. For starters, AI does not 'learn' like a human learns. But still many of the main terms of the field are based on learning: terms like 'learning rate', 'neural networks', or 'deep learning' are implying that there's some kind of being which learns, not just a very complicated decision tree. It's not all the fault of hype marketing people!
> AI-scientists themselves are partially to blame for this

They are not addressing the public or swaying opinion

I disagree; consider the use of the term "video game AI", which historically at least has just been a bunch of _if_ statements chained together. This is totally valid, it's an example of AI without machine learning.

The thing is that AI is just about the most general term for the type of computing that gives the illusion of intelligence. Machine learning is a more specific region of the space of AI, and generally is made of statistical models that lead to algorithms that can train and modify their behavior based on data. But this includes "mundane" algorithms like k-means clustering or line-fitting. Deep learning (aka neural networks) is yet a more specific subfield of ML.

I think the term AI just has more "sex appeal" because people confuse it with the concept of AGI, which is the holy grail of machine intelligence. But we don't even know if this is achievable, or what technology it will use.

So in terms of conceptual spaces, we can say that AI > ML > DL, and we can say (by definition) that AI > AGI. And it seems very likely that AGI > ML. But it's not known, for instance, whether AGI > DL, ie, we don't know for sure that deep learning/neural networks are sufficient to obtain AGI.

In any case, people should put less weight on the term AI, as it's a pretty low bar. But also yes, the term is way over hyped.

I'm thinking of cases such as colleagues selling as "ML" something they were then forced to admit as "we use SQL to pick out instances of this specific behaviour we knew was happening". Embarrassing all round.

As folks that work in tech we can tell the difference between stuff that's got some form of depth to it in "proper" AI: ML, DL, AGI as you suggest, vs the over-hyped basic computation stuff. And the selling of the latter as the former can rankle.

I'm working on a project that uses GPT-3 and similar stuff, even before the hype. I think the overhype is really tiring.

Just like with most of these hype cycles there is an actual useful interesting technology, but the hype beasts take it way overboard and present it as if it's the holy grail or whatever. It's not.

That's tiring, and really annoying.

It's incredibly cool technology, it is great at certain use cases, but those use cases are somewhat limited. In case of GPT-3 it's good at generative writing, summarization, information search and extraction, and similar things.

It also has plenty of issues and limitations. Lets just be realistic about it, apply it where it works, and let everything else be. Now it's becoming a joke.

Also, a lot of products I've seen in the space are really really bad and I'm kinda worried AI will get a scam/shitty product connotation.

> GPT-3 it's good at generative writing

made up bullshit

> summarization

except you can't possibly know the output has any relation whatsoever to the text being summarized

> information search and extraction

except you can't possibly know the output has any relation whatsoever to the information being extracted

people still fall for this crap?

Agreed. Been testing out responses to parsing complex genomics papers (say, a methodology section describing parameters of some algorithm) and its mostly rephrasing rather than digesting and responding with useful information / interpretation. And it will use so many words and imbue so little to the conversation, yet appear like it's helping because ... words.
> complex genomics papers

I think it's fair to say this is not one of the use cases where it shines. It's not great at logic, it's also not that smart.

That's exactly what the hype does. Too big claims and then it gets dismissed when it inevitably doesn't live up to the hype.

Yeah you can't use it for academic research, it's really, really terrible at it.

It will even claim it can generate citations for you too, which is pretty messed up because when I tried it just fabricated them replete with faked DOIs.

Where it shines is at squishy language stuff, like generating the framework of an email, paraphrasing a paragraph for you, or summarizing a news article.

It really is revolutionary at language tasks, but unfortunately the hype machine and these weird "AI sycophants" have caused people to dramatically overestimate it's use cases.

I think this is sort of the other side of the hype, totally dismissing it is also incorrect imo.

Yes, it's overhyped, but it's not useless, it actually does work quite well if you apply it to the right use cases in a correct way.

In terms of accuracy, in ChatGPT the hallucination issue is quite bad, for GPT3 it's a lot less and you can reduce it even further by good prompt writing, fine tuning, and settings.

Can we just recognize it for what it is?

Someone called it a zeroday on human cognition, on the entire society so I am ready to recognize for that.
> I'm kinda worried AI will get a scam/shitty product connotation

I think we're already there. A legion of AI based startups seem to be coming out daily (https://www.futuretools.io/) that offer little more than gimmicks.

You are probably right, kinda sad.

My last resort is to just remove all AI references from my marketing and just deliver the product.

> Just like with most of these hype cycles there is an actual useful interesting technology, but the hype beasts take it way overboard and present it as if it's the holy grail or whatever. It's not.

See also: Gartner hype cycle

Thanks! Very interesting. I guess our challenge will be surviving the inevitable dip.
> Also, a lot of products I've seen in the space are really really bad and I'm kinda worried AI will get a scam/shitty product connotation.

I agree with this, I feel like I've seen a lot of really cool technology get swept up in a hype storm and get carried away into oblivion.

I wonder what ways there are for the people who put out these innovations to shield them/their products from it?

Luckily I have a lot of faith in the OpenAI people - I hope their shielding themselves from the technological form of audience capture.

Finally, a take on chatgpt and similar LLMs I agree with!

I've criticized it whenever it gets brought up as an alternative for academic research, coding, math, other more sophisticated knowledge based stuff. In my experience at least, it falls apart at reliably dealing with these and I haven't gone back.

But man, is it ever revolutionary at actually dealing with language and text.

As an example, I have a bunch of boring drama going on right now with my family, endless fucking emails while I'm trying to work.

I just paste them into chat gpt and get it to summarize them, and then I get it to write a response. The onerous safeguards make it so I don't have to worry about it being a dick.

Family has texted me about how kind and diplomatic I'm being and I honestly don't even really know what they're squabbling about, it's so nice!

Haha that's amazing. This is exactly what I mean, for the right use cases it's absolutely amazing.

Good luck with the drama! Make sure to read a summary for the next family meeting haha.

The great part is I can just get it to summarize all the summaries, love how it's flexible like that.

Yeah I will be sure to read it before meeting them, would be awkward if they found out I was using it during one of the disputes, which was whether or not to keep resuscitating Grandma.

ChatGPTs stupid ethical filters made it so I actually had to type my response to that one all by myself.

I think the hope is that unlike crypto, as others have said, clearly AI will have actual good applications, so the hope is the hype beasts, after they've burned through all the grifting they can, they can go off back to selling penis pills or whatever they were into last year (or may be pre 2020).
> I'm kinda worried AI will get a scam/shitty product connotation.

Which has happened before. The original semantic/heuristic AI, most notably expert systems, over-promised and ultimately under-delivered. This led directly to the so-called "AI winter" which lasted more than two decades and didn't end until quite recently. It's a very real concern, especially among people who want to push the technology forward and not just profit from it.

Yes and it is sad because I see people trying to find problems for ChatGPT and say this is the future.. but the industries they are targeting have such a wide variety of concrete problems for a startup to solve.
Absolutely not.

It seems too exciting to me and I am eager to see more AI. It's fascinating stuff.

Happy to find a fellow soul. I'm fatigued of the complaints of AI fatigue - especially where the complaints aren't based on recent (last year or so) first hand use.

It's bold (to put kindly) how lengthy some of these critical comments are from folks who later in the thread admit to not personally used Copilot (for example) much themselves.

The quality of LLM output can wildly vary based on what prompts (or series of prompts) are used.

Well most products today need to be adaptive and handle complex states, even if there's a few if sentences in there handling something in an intelligent way that's AI, it's a broad term after all. The problem is that it's also become a marketing buzzword.
I work in the field, so: yes, since about 2015, heh.
That’s Capitalism. AI is the new growth frontier, so it’s all you are hearing about. Seems like LLMs and generators are genuine innovations. But don’t lose sight that these innovations are driven by the Capitalist need to concentrate more surplus value into fewer hands. This is no different than programable looms, etc. of the past, except now they will try to automate immaterial/“intellectual” work. It remains to be seen if these technologies will succeed at that, but the Capitalists are compelled to try, and we will be forced to live with the wreckage.
The "I" in AI is just complete bullshit and I can't understand why so many people are in a awe of a bit of software that chains words to another based on some statistical model.

The sad truth is that ChatGPT is about as good an AI as ELIZA was in 1966, it's just better (granted: much better) at hiding its total lack of actual human understanding. It's nothing more than an expensive parlor trick, IMHO.

Github CoPilot? Great, now I have to perform the most mentally taxing part of developing software, namely understanding other people's code (or my own from 6 months ago...) while writing new code. I'm beyond thrilled ...

So, no, I don't have an AI fatigue, because we absolutely have no AI anywhere. But I have a massive bullshit and hype fatigue that is getting worse all the time.

I can say with a certain degree of confidence that you haven't actually used CoPilot daily.
This is such a bullshit answer. No, I don't use it daily because I tried it for a couple hours and it suggested nothing useful and several harmful. Why would I keep using it?
[flagged]
I've used it and it did nothing helpful. I also find autocomplete slows me down. The code it suggested always needed enough reworking I would have been faster writing it out from scratch. It's just not that helpful for me. Maybe if I didn't know the apis that well but I suspect even then it would be as much a liability as a benefit
> They replace the default autocomplete in a way which largely unnoticeable, but surprisingly effective at complex autocomplete tasks.

I've yet to see it. It's barely above IDEAs autocomplete in the rare cases when it manages to trigger on my code, and it has already been wrong more than once in the few times it did deign to provide autocompletion.

Hey, maybe don't outright call people liars about their own lived experiences just because they don't agree with yours...?
90% of the time CoPilot-bros stop at this point without giving any good examples of how this post-autocomplete monster helped them. Autocomplete works quite well in most use cases - it is low effort, free and most importantly, ethical. CoPilot on the other hand jumps through so many hoops to generate something marginally and arguably better, but at the cost of what? This is exactly like the Search Vs ChatGPT problem - do you want a deterministic, algorithmic, fine tuned experience or some random probabilistic, overconfident crap.
I don't find basic autocomplete useless, although I probably use it less than most people. (For example, I rarely write Java and when I do I don't use long names or deeply nested structures and I don't implement equals etc. by default.) I think people use autocomplete in two ways; when I use it I know what I want to be in the code and it's a way to type 30 characters by pressing 3 keys or whatever. But I also see a lot of people use it like "I don't know what to do next, what methods are available?" And this is usually to the detriment of the code quality.

Copilot is not like autocomplete. It only works in the second mode, because any nontrivial code it generates needs to be read, considered, and understood. (And any trivial code it generates can be done by autocomplete or long-existing non-AI tools.) This is especially true given LLMs' hallucinatory behavior - by definition it will often spit out something that "looks right" even if it's absolutely not - and such code is harder to review than code that looks obviously wrong.

So if you do use autocomplete in the second mode, maybe you find Copilot a super-powered version of that. And if you have the same weaknesses as Copilot, reviewing its code after it's done writing it is probably not any different than reviewing your own code after writing it, so for you it takes the same amount of time. For me, that's not the case.

When I used it, and when I see others use it, Copilot is like an impossibly overenthusiastic junior developer I will never be able to teach better habits to.

I've worked with teams that used Copilot. They claim it's great "Hey, now I don't have to actually spend any time writing all this boilerplate!" while for me, the person who has to review their code before releasing stuff, easier ways of writing boilerplate is not a positive, it's a negative.

If writing boilerplate becomes effortless, then you'll write more of it, instead of feeling the pain of writing it and then trying to reduce it, because you don't want to spend time writing it.

And since Copilot was accepted as a way to help the developers on the teams, the increase of boilerplate have been immersive.

I'm borderline pissed, but mostly at our own development processes, not at Copilot per se. But damn if I didn't wish it existed somehow, although it was inevitable it would at one point.

>the increase of boilerplate have been immersive

Has it really? Or are you worried that this is something that will happen?

Of course I don't know how other people use it but I find that it's very much like having a fairly skilled pair programmer on board. I still need to do a lot of work but I get genuine help. I don't find that I personally write more boilerplate code than before, every programming principle applies as it always has.

I wrote it in past tense, it's based on actual situations :) If you don't believe what I write, I guess it doesn't matter what I write now. Regardless.

One simple example that I've had to reject more than once.

- Function 1 does something

- Developer needs something like Function 1 but minor change

- Developer starts typing name of function which has a similar name to Function 1, but again, minor difference

- Copilot helpfully suggests copy-pasting Function 1 but with the small change incorporated

- Developer accepts it, commits and sends the patch my way

Rather than extracting the common behavior into it's own function and call that from both of them, refactors which Copilot doesn't suggest, the developers is fine with just copy-pasting the function.

Now we have to maintain two full slightly different functions, rather than 1 full functions + 2 minor ones.

Obviously a small example, and it wouldn't be worth extracting it the first time it happens or on a smaller scale. But once you have entire teams doing something like this, it becomes a bit harder to justify copy-paste approach, especially when you want the codebase not to evolve to complete spaghetti.

And finally, I'm not blaming the tool, it's not Copilots fault. But it does seem to have made developers who rely on it think less, compared to the ones that don't.

I feel ya. If your job is to kick back bad code, and now there is a tool that generates bad code, how does this not make your job more important?

Why not get some of the freed up, Copilot augmented developer labor budget moved to testing and do more there or build more tools to make your personal, boilerplate, repetitive tasks more efficient?

If the coders are truly just dumping bad code your way, that's an externality and the cost should be called out.

I use github copilot on a daily basis and it improves my time from thinking to code.

Often I have times where I'm think about a specific piece of code that I need and I have it partially in my head and github copilot "just completes" it. I press tab and that's it.

I'm not talking about writing entire functions where you have to mentally strain yourself to understand what it wrote.

But I've never seen any autocompleter do it so good then github copilot. Even for documentation purposes like JSdoc and related commenting system it's amazing.

It's a tool I pay for now since it's proven to be a tool that increases my productivity.

Is it gonna replace us? I hope not, but it does look promising as one of those tools people will talk about in the future.

It would be helpful if people could include in their assessment roughly how much time they've personally spent using these tools.

Helping write boilerplate is to Copilot what cropping is to Photoshop.

Some of the ways I've found Copilot a powerful tool in my toolbox: Writing missing comments (especially unfamiliar code bases), "translating" parts of unfamiliar code to a more familiar language, suggesting ideas for how to implement a feature (!) in comments.

you could stop doing code reviews and do something else
Yep. I’m personally skeptical of so many other use cases for LLMs but CoPilot is fantastic and basically just autocomplete on rocket fuel. If you can use autocomplete, you can use CoPilot super effectively.
I almost always turn autocomplete off except in circumstances where the API has bad documentation. I also found that copilot was an aggravation more than a help after using it for a couple weeks.
We programmers enjoy writing code. We derive satisfaction when a code is perfect and elegant. But its going to end very soon. Artists are freaking out because things that takes them days to create now only take 2 seconds. We are next.

The writing is on the wall. Programming as we know it is going to end. We should be embracing these tools and should start moving from software developers to software architects role.

I don't think you're getting what I'm saying. I'm _faster_ without autocomplete
I've used it quite a lot and I agree with the original post. It seemed really useful at first but then it started introducing several bugs in large blocks of code. I've stopped using it in the end since the small snippets on the one line size is trivial enough to write myself (with just vim proficiency) and the larger blocks on the order of a function autocomplete is too bug prone (and kills too much willpower budget to fix).
I haven't. Now you know for a fact :)

What I have seen about it ranged from things that can be nearly just as well handled by your $EDITOR's snippet functionality to things where my argument kicked in - I have to verify this generated code does what I want, ergo I have to read and understand something not written by me. Paired with the at least somewhat legally and ethical questionable source of the training data, this is not for me.

So stop evangelizing about stuff you haven’t used. Understanding code is easier than writing it from the scratch. That’s why code review doesn’t take as much time as writing code and you still need to prove your code works, even if you wrote it yourself.
I am not evangelizing, I am just stating why this is not for me and my way to write software.
Understanding code is only easier for simple tasks. I've definitely had copilot spit out complex algorithms that looked right at first glance but actually had major issues that required me to write it from scratch.
Thats why you test, this could also happen with code you wrote so it’s not an argument against copilot. Did you wrote your “complex algorithm” and then run and debugged it in your head? No, you’ve tested it. Do the same with Copilots code
If testing is the equalizer, there is no difference between black box code and something you fully understand. Which fair enough, is how ML works in general.
I contend understanding the semantics of code is harder than writing the syntax. Reading the syntax without thinking deeply (to the level needed to write it, or deeper) seldom helps you realize unexpected corner cases. This is why stochastic testing is so valuable.

ad.: Code review takes less time than writing code for the same reason reading a book takes less time than writing one. Distillation and organization of ideas requires expertise gained through experience and long thought. Reading a book requires reading ability.

Understanding a book (and the intricacies underlying it) takes effort on the order of the original writing, but most people don't seek that level of understanding. The same is true of code.

Code review often takes me longer than writing code. More generally, reading other people's code is more difficult for me than writing (or reading) my own.
I can say with a higher degree of confidence that you haven't actually used CoPilot daily for any respectably sized project.
“AI” isn’t bull shit, it’s correctly labeled. It’s intelligence which is artificial: i.e. fake, ersatz, specious, not genuine… It’s our fault for not just reading the label. (I absolutely agree with your post and your viewpoint, just to be clear!)
"Artificial" is not synonymous with "fake". "Fake" implies a level of deception.
Not necessarily true. People talk about “fake meat” all the time but it’s clear there’s no level of fraudulence implied by this usage. It’s meant in the sense of “artificial meat”. There are multiple ways the word “fake” is used, and one is as a synonym for “artificial”.

However, in this case, it does seem that there is a level of fraudulence and deception. Given that “fake” often is used exactly the way you say, maybe “fake intelligence” would indeed be a more appropriate term.

Fake meat is still a form of deception; it's something that's not meat pretending to be meat. If lab grown meat gets good enough to be indistinguishable from "real" meat, then it would no longer be fake, it would just be artificial.
Is “fake grass” (i.e. astroturf) a form of deception in your eyes?
When seen at a distance, it can trick people into thinking you have a perfectly manicured lawn. But you're just lazy/evil/it's a holiday house. Deception!
Have you been around people who say "fake meat?" Every time I've heard it, it was said derisively and implied fraudulent meat.
Yes. I’ve heard loads of people refer to it as fake without an implied pejorative meaning.
The people I know that like it use artificial and the people that won't try it use fake.
Fascinating. Not all people are the same. Who’d have thought?
Definitely. Quirks like this are also what makes AI difficult.
Superficial Intelligence may hit the mark there.
Artifical means "not human" in this context for me, but I understand "Intelligence" as the abiltiy to actual reason about something based on things you learned and/or experienced, and these "AI" tools don't do this at all.

But defining "intelligence" is a philosopical question that doesn't necessarily have one answer for everything and everyone.

Personally, I try to take a more inductive approach. We don’t know what intelligence is, but we assume it’s something we exhibit. We also clearly recognize other animals as possessing the same trait to varying degrees. Since we don’t know what it is, and since (I would argue) we can only convincingly claim that exists in other biological organisms without meeting a high burden of proof, to claim that it exists in an inorganic substrate requires a VERY large burden of proof to be met, similar to what would be met if you were claiming that magic existed. In my view, calling computers “intelligent” is in the same league as claiming that crystals are magic. Of course, this depends on my own philosophical interpretation of what intelligence is, as you say.
Intelligence is a capability not a mechanism, and therefore if you're able/willing to define what that capability is, there should be no problem measuring/gauging the intelligence of any system, biological or not. You don't need to look inside the black box - you only need to test if the black box has this capability.

Intelligence may be a fuzzily defined word in everyday usage, but I don't think it's the mystery you present it to be. Joe public may argue against any and all definitions of the word that they personally disagree with (maybe just dislike), but it's nonetheless quite easy to come up with a straightforward and reductive definition if you actually want to!

You and I are clearly referring to two different things when we use the word “intelligence”. It is also not nearly so easy to come up with a simple/mechanical/verifiable definition for the thing that you’re referring to. Unless you have a good definition—in which case, you might want to put your money where your mouth is and get busy revolutionizing multiple fields of human inquiry!

It’s also plain that many people are very interested in looking inside the black box and think the contents of the black box are relevant and important. This fact doesn’t change just by your saying so.

People are interested in looking inside the black box (our brain) for sure, partly for inspiration as to how to implement intelligence among other things, but implementing isn't the same as defining.

Being able to define what you want to achieve isn't generally the same as knowing HOW to achieve it (except in this case the definition of intelligence rather does suggest the right path).

The intention of the "artificial" in "AI" is not that particular meaning of "artificial", but the one for "constructed, man-made"—see meaning #1 in the Wiktionary definition[0]; the one you are using is #2.

It is often frustrating that English has words with such different (but clearly related) definitions, as it can make it far too easy to end up talking past each other.

[0] https://en.wiktionary.org/wiki/artificial

[flagged]
We've banned this account for repeatedly breaking the site guidelines and ignoring our request to stop.

If you don't want to be banned, you're welcome to email hn@ycombinator.com and give us reason to believe that you'll follow the rules in the future. They're here: https://news.ycombinator.com/newsguidelines.html.

As much as I’m sick of AI products, I’m even more sick of the “ChatGPT is bullshit” argument.
I like this take. It has many clear applications already and LLM's are still only in their infancy. I both criticize and use ChatGPT at work. It has flaws and it has advantages. That it's bullshit or "ELIZA" is a short-sighted view that overvalues the importance of AGI and misses what we're already getting.

But yes indeed, there are many, many AI products launched during this era of rapid progress. Even kind of shoddy products can be monetized if they provide value over what we had before. I think the crowded market and all the bullshit and all the awesome, all at once, is a sign of very rapid progress in this space. It will probably not always be like this and who knows what we are approaching.

How are you using it at work?
I've used it to proof emails for grammar, and it's done ok.

I'll also throw random programming questions into it, and it's been hit and miss. SO is probably still faster, and I like seeing the discussion. The problem with chatGPT right now is it gives an answer like it's certainty when it's often wrong.

I can see the benefits of this interaction model (basically summarizing all the things from a search into what feels like a person talking back), but I don't see change the world level hype at the moment.

I also wonder if LLMs will get worse over time through propagation error as content is generated by other LLMs.

I’m not the person you replied to but I’ve been using OpenAI’s API a lot for work. Some examples:

- Embedding free text data on safety observations, clustering them together, using text completion to automatically label the clusters, and identifying trends

- Embedding free text data on equipment failures. Some of our equipment failures have been classified manually by humans into various categories. I use the embeddings to train a model to predict those categories for uncategorized failures.

- Analyzing employee development goals and locating common themes. Then using this to identify where there are gaps we can fill in training offerings.

The only problem with the “ChatGPT is bullshit” argument is that it is only half true.

ChatGPT, when provided with a synthetic prompt is reliably a synthesizer, or to use the loaded term, a bullshiter.

When provided with an analytic prompt, it is reliably a translator.

Terms, etc: https://www.williamcotton.com/articles/chatgpt-and-the-analy...

> ChatGPT, when provided with a synthetic prompt is reliably a synthesizer, or to use the loaded term, a bullshiter.

sounds like most people tbf

There are people who in many situations use as much critical though as ChatGPT does.

ChatGPT isn't as good as a human who puts in a lot of effort, but in many jobs it can easily outperform humans who don't care very much.

(comment deleted)
It can be both bullshit and utterly astounding.

In terms of closing the gap between AI hype and useful general purpose AI tools, no one can reasonably deny that it's an absolute quantum leap.

It's just not a daily driver for technical experts yet.

> quantum leap

Ironically accurate.

In normal English usage, a quantum leap is a step-change, a near-discrete rather than continuous improvement, a large singular advance.

Given we are not talking about state changes in electrons, there is nothing wrong with this description of ChatGPT - it truly does feel like a massive advance to anyone who has even cursorily played with it.

For example, you can ask it questions like "Who was born first, Margaret Thatcher or George Bush?" and "Who was born first, Tony Blair or George Bush?" and in each instance it infers which George Bush you are talking about.

I honestly couldn't imagine something like this being this good only three years ago.

(1) You are correct in that placing both of those questions into Google doesn't quite get you anywhere near the answer that I imagine ChatGPT gives you (as you point out). Although, Google does "infer" which Bush you are talking about, there isn't a clear "this person is older" answer, you have to dive into the wiki pages basically to get the answer.

(2) Counter. I asked it the other day "how many movies were Tom Hanks and Meg Ryan in together" and the answer ChatGPT gave was 2 ... not only is that wrong it is astonishingly wrong (IMO). You could be forgiven for forgetting Ithaca from 2015. I could forgive ChatGPT for forgetting that one. But You've Got Mail? That's a very odd omission. So much so I'm genuinely curious how it could possible get the answer wrong in that way. And for the record, Google presents the correct answer (4) in a cut out segment right at the top, a result and presentation very close to what one would expect from ChatGPT.

I don't know about other use cases like generating stories (or tangentially art of any kind) for inspiration, etc. But as a search engine things like ChatGPT NEED to have attributions. If I ask the question "Does a submarine appear in the movie Battlefield Earth?" it will confidently answer "no". I _think_ that answer is right, but I'm not really all that confident it is right. It needs to present the reasons it thinks that is right. Something like "No. I believe this because (1) the keyword submarine doesn't appear in the IMDb keywords (<source>), (2) the word submarine doesn't appear in the wikipedia plot synopsis (<source>), (3) the film takes place in Denver (<source>) which is landlocked making it unlikely a submarine would be found in that location during the course of the film."

The Tom Hanks / Meg Ryan question/answer would at least more interesting if it explained how it managed to be so uniquely incorrect. That question will haunt me though ... there's some rule about this right? Asking about something you have above average knowledge in and watching someone confidently answer it incorrectly. How am I supposed to ever trust ChatGPT again about movie queries?

The biggest thing I’ve learned from chatGPT is that real people struggle with the difference between intelligence, understanding, and consciousness / sentience.
Because they are all ill defined in the manner they are used in common language. Hell, we have trouble describing what they are, especially in a scientific fact based setting.

Before this point in history we accepted 'I am that I am' because there wasn't any challenger to the title. Now that we are putting this to question we realize our definitions may not work well.

Which should be no surprise, as people have been grappling with these ideas for centuries, and we still don't have any definitive idea of what consciousness/sentience truly is. What I find interesting is that at one point the Turing test seemed to be the gold standard for intelligence, but chatGPT could pass that with flying colors. So how exactly will we know if/when true intelligence does emerge?
Well, my point wasn’t that there is a good definition of consciousness.

My point was that “consciousness” and “intelligence” are very different things. One does not imply the other.

Consciousness is about self reflection. Intelligence is about insight and/or problem solving. The two are often correlated, especially in animals, especially in humans, but they’re not the same thing at all.

“Is chatgpt consciousness” is a totally different question than “is chatgpt intelligent”.

We will know chatgpt is intelligent when it passes our tests of intelligence, which are imperfect but at least directionally correct.

I have no idea if/when we we know whether chatgpt is conscious, because we don’t really have good definitions of consciousness, let along tests, as you note.

The most annoying thing to me is people thinking AI wants things and gets happy and sad. It doesn't have a mamailian or reptilian brain. It just holds a mirror up to humanity generally via matrix math and probability.
Well said. It is a mistake to anthropomorphize large language models; they really hate that.
>The biggest thing I’ve learned from chatGPT is that real people struggle with the difference between intelligence, understanding, and consciousness / sentience.

Well, I'm no fan of chatGPT. But it appears most people are worse than chatGPT, because just regurgitate what they hear with no thought or contemplation. So you can't really blame average folks who struggle with the concepts of intelligence/understanding that you mention.

I agree with you completely. I work in the field and I think your sentiment is way more common amongst people who know about the technology, vs the fair weather fans who have all jumped on the hype bandwagon recently. I actually posted the same thing (that it's no different than Eliza) a month or so ago, and got at least one hilarious dismissal, like the "I bet you make widgets" person that replied to you.
If you believe that ChatGPT is similar to Eliza, then I can guarantee that you have no rigorous no-wriggle-room definition of what intelligence is. Maybe you think you understand it, or have defined it, but I'm 100% certain any such definition is not 100% reductive and instead relies on other ill-defined works like "reasoning" etc etc.
> Github CoPilot? Great, now I have to perform the most mentally taxing part of developing software, namely understanding other people's code (or my own from 6 months ago...) while writing new code. I'm beyond thrilled ...

I think there's an argument to be made that AI is being used here to help you tackle the more trivial tasks so you have more time to focus on the more important, and challenging tasks. Albeit I recognise GitHub CoPilot is legally questionable.

But yes, I agree with your overall point that AI has still not been able to 'think' like a human but rather can only still pretend to think like a human, and history has shown that users are often fooled by this.

I think the parent’s comment is probably referring to the fact if you use Copilot to write code then you have to go through and try to understand what it wrote and possibly debug it. And you don’t have the opportunity to ask it why it wrote it the way it did when reviewing its code.
I think you’re right, but that just means parent doesn’t understand copilot and is off tilting at windmills.

Copilot is amazing for reducing the tedium of typing obvious but lengthy code (and strings!). And it’s inline and passive; it’s not like you go edit -> insert -> copilot function and it dumps in 100 lines of code you have to debug. Which is what it sounds like parent is mistaking it for.

I’m reminded of 1995, when an elderly relative told me everything wrong with the internet based on TV news and not having ever actually seen the internet.

> Copilot is amazing for reducing the tedium of typing obvious but lengthy code (and strings!)

Which it occasionally mistypes. Then you're off to chase a small piece of error in a tub of boilerplate. Great stuff! For actual example, see [0]

[0] https://blog.ploeh.dk/2022/12/05/github-copilot-preliminary-...

You must be a much better programmer than I if those are examples you’d use copilot for. I was thinking more like:

   start_value = get_*start_value(user_input)*
   self.log.d*ebug(‘got start_value {start_value}’)*
. . . where the would-be italics are what copilot would likely suggest for completion.

And if it’s wrong, you just. . . keep typing. It’s autocomplete, just like IDEs have for other things. I’m kind of astounded that people have such an emotional reaction to an optional, low-key, passive, easily-ignored tool that sometimes saves a bunch of typing. Yes, if you always accept the suggestions you’ll have problems. Just like literally every other coding assistance tool.

That's not my blog, I just thought the example to be relevant.

> I was thinking more like:

That example is straight up from any of those "programming is not bound by typing speed" essays of yore.

> people have such an emotional reaction to an optional, low-key, passive, easily-ignored tool that sometimes saves a bunch of typing.

Maybe because it's not generally advertised by proponents as "an optional, low-key, passive, easily-ignored tool that sometimes saves a bunch of typing"? Just look at the rest of the thread, it's pronounced as a game-changer in productivity.

Different experiences, I guess. I’m a low end, part-time hobbyist programmer, and for me at least 75% of my time is spent essentially typing in obvious, easily-checked code. It has been a game changer for me. It’s also led me to write better comments, because rather than being a pure tax, they improve the generated code.

I can see how someone who’s always working on sophisticated, mentally challenging code would get less benefit and would see more frequent errors.

But it's trickling in small chunks at a time unless you are just smashing tab repeatedly and don't look at what it did until the very end. You can also not accept what it offers and just continue writing code for yourself. If a dev submitted a bunch of Copilot code they don't understand and can't answer questions about you reject the PR outright and they eventually realize it didn't save them any time or effort. Copilot isn't the employee.
As soon as I open a fresh IDE these days I immediately miss CoPilot and it's the first thing I install.

Hype or not, it's incredibly useful and has increased my productivity by at least 20%. Worth every penny.

ChatGPT is of actual help for me in various daily tasks, which was never the case with ELIZA or earlier chatbots which were only good as a curiosity or to have some fun.

Lack of actual human understanding? Of course, by definition a machine will always lack human understanding. Why does that matter so much if it's a helpful tool?

For what it's worth, I do agree that there is a lot of hype. But contrary to blockchain, NFTs, web3, etc., this is actually useful for many people in many everyday use cases.

I see it as more similar to the dot com hype - buying a domain and creating a silly generic website didn't really multiply the value of your company as some people thought in that era, but that doesn't mean that websites weren't a useful technology with staying power, as time has shown.

(comment deleted)
“It’s just statistics” is an evergreen way to dismiss AI. The problem is you’re also just statistics.
Shh. The models don’t like hearing that.
Source for consciousness / intelligence to be "statistics"?

I don't think there is any because there is no functional model for what organic intelligence is or how it operates. There are plethora of fascinating attempts / models but only a subset implore that it is solely "statistical". And even if it was statistical, the implementation of the wet system is absolutely not like a gigantic list of vectorized (stripped of their essence) tokens

There’s a man who claims to have solved consciousness as a multilayered Bayesian prediction system.

See Scott Alexander for attempts to explain what is apparently impenetrable papers on it.

Sources for intelligence to be magic? I mean we know it's complicated but intelligence also spans the smallest creatures on the planet to humans. This points at intelligence being a reduceable problem that is layered. On top of that it's unlikely we need to model nerve behavior to get something intelligence like output.
That's like saying that airplanes aren't flying since they're not flapping their wings. Intelligence is a capability - not a specific mechanism.

Consciousness is a subjective experience (regardless of what you believe/understand to be responsible for that experience), so discussing "consciousness/intelligence" is rather like discussing "cabbages/automobiles".

(comment deleted)
The "I" in AI is just complete bullshit

We're about six minutes away from "AI bros" becoming a thing.

The same kind of grifters who always latch onto the latest thing and hype it up in order to make a quick buck are already knocking on AI's door.

See also: Cryptocurrency, and Beanie Babies.

I'm more fatigued by people denying the obvious that ChatGPT and similar models are revolutionary. People have been fantasizing about the dawn of AI for almost a century and none managed to predict the rampant denialism of the past few months.

I suppose it makes sense though. Denial is the default response when we face threats to our identity and sense of self worth.

So, to you, ChatGPT is approaching AGI?
I do believe if we are going to get AGI without some random revolutionary breakthrough, to achieve it iteratively, It's going to come through language models.

Think about it.

What's the most expressive medium we have which is also absolutely inundated with data?

To broadly be able to predict human speech you need to broadly be able to predict the human mind. To broadly predict a human mind requires you build a model of it, and to have a model of a human mind? Welcome to general intelligence.

We won't realize we've created an AGI until someone makes a text model, starts throwing random problems at it, and discovers that it's able to solve them.

> I do believe if we are going to get AGI without some random revolutionary breakthrough, to achieve it iteratively, It's going to come through language models.

Language is way, way far removed from intelligence. This is well-known in cognitive psychology. You'll find plenty of examples of stroke victims who are still intelligent but have lost the ability to produce coherent sentences, and (though much rarer) examples of people who can produce clear, eloquent prose, yet are so learning and mentally challenged that they can't even tell the difference between fantasy and reality.

We don't judge AI by their ability to produce language, we judge them by their conference and ability to respond intelligently, to give us information we can use.
> To broadly be able to predict human speech you need to broadly be able to predict the human mind

This is a non sequitur. The human mind does a whole lot more than string words together. Being able to predict which word would logically follow another does not require the ability to predict anything other than just that.

I think what the commenter is saying is that, in time, language models too will do a lot more than string words together. If it's large enough, and you train it well enough to respond to “what's the best next move in this chess position?” prompts with good moves, it will inevitably learn chess.
I don't think that follows, necessarily. Chess has an unfathomable amount of states. While the LLM might be able to play chess competently, I would not say it has learned chess unless it is able to judge the relative strength of various moves. From my understanding, an LLM will not judge future states of a chess game when responding to such a prompt. Without that ability, it's no different than someone receiving anal bead communications from Magnus Carlsen.
An LLM could theoretically create a model with which to understand chess and predict a next move, you just need to adjust the training data and train the model until that behavior appears.

The expressiveness of language lets this be true of almost everything.

Exactly. Since language is a compressed and transmittable result of our thought, to predict text as accurately as possible requires you do the same. A model with understanding of the human mind will outperform one without.

> Being able to predict which word would logically follow another does not require the ability to predict anything other than just that.

Why? Wouldn't you expect that technique to generally fail if it isn't intelligent enough to know what's happening in the sentence?

"The ability to speak does not make you intelligent." — Qui-Gon Jinn, The Phantom Menace.
Why the obsession with AGI? The point is that ChatGPT is already useful.
Is it? I see it mostly generates BS much faster.
Brothers Grimm would like a word with you about what "BS" means.

ChatGPT is good at making up stories.

Perhaps a more interesting question is "how much better do we understand what characteristics AGI will have due to ChatGPT?"

We don't really understand what intelligence means -- in humans or our creations -- but ChatGPT gives us a little more insight (just like ELIZA, and the psychological research behind it, did).

At the very least, ChatGPT helps us build increasingly better Turing tests.

Yes. It is obviously already weak AGI (obvious to anyone if they saw it 20 years ago).

It is also obvious that we are in the middle of a shift of some kind. Very hard to see from within, but clearly we will look back at 2022 as the beginning of something

> I suppose it makes sense though. Denial is the default response when we face threats to our identity and sense of self worth.

It's important to note that this is your assumption which I believe to be wrong (for most people here).

There's a fellow that kinda predicted it in 1950 [0]:

> These arguments take the form, "I grant you that you can make machines do all the things you have mentioned but you will never be able to make one to do X."

> [...]

> The criticisms that we are considering here are often disguised forms of the argument from consciousness, Usually if one maintains that a machine can do one of these things, and describes the kind of method that the machine could use, one will not make much of an impression.

Every time "learning machines" are able to do a new thing, there's a "wait, it is just mechanical, _real_ intelligence is the goalpost".

[0] https://www.espace-turing.fr/IMG/pdf/Computing_Machinery_and...

>Every time "learning machines" are able to do a new thing, there's a "wait, it is just mechanical, _real_ intelligence is the goalpost".

Just because people shift the goalposts doesn't mean that the new position of the goalposts isn't closer to being correct than the old position. You can criticise the people for being inconsistent or failing to anticipate certain developments, but that doesn't tell you anything about where the goalposts should be.

> ChatGPT and similar models are revolutionary

For _what purpose_, tho? It's a good party trick, but its tendency to be confidently wrong makes using it for anything important a bit fraught.

If you're the type of person that struggles to ramp up production of a knowledge product, but has great success in improving a knowledge product through an iterative review process, then these generative pre-trained transformers are fantastic tools in your toolbox.

That's about the only purpose I've found so far, but it seems a big one?

If you work at a computer, it will increase your productivity. Revolutionary is not the word I'd use, but finding use cases isn't hard.
I can buy that it's a better/worse search engine (better in that it's easier to formulate a query and you get the response right there without having to parse the results; worse in that there's a decent chance the response is nonsense, and it's very confident when it's being wrong about things).

I can't really imagine asking it a question about anything I cared about and not verifying via a second source, though, given its accuracy issues. This makes it feel a lot less useful.

(comment deleted)
But will it? After accounting for the time needed to fix all the bugs it introduces?
Humans introduce bugs too. ChatGPT is still new, so it probably makes more mistakes than a human at the moment, but it's only a matter of time until someone creates the first language model that will measurably outperform humans in this regard (and several other important regards).
>> it's only a matter of time

That reminds me how in my youth many were planning on vacations to Mars resorts and unlimited fusion energy) Stars looked so close, only a matter of time!

> it's only a matter of time until someone creates the first language model that will measurably outperform humans in this regard

This seems to have been the rallying cry of AI-ish stuff for the past 30 years, tho. At a certain point you have to ask "but how much time"? Like, a lot of people were confidently predicting speech recognition as good as a human's from the 90s on, for instance. It's 2023, and the state of the art in speech recognition is a fair bit better than Dragon Dictate in the 90s, but you still wouldn't trust it for anything important.

That's not to say AI is useless, but historically there's been a strong tendency to say, of AI-ish things "it's 95% of the way there, how hard could the last 5% be?" The answer appears to be "quite hard, actually", based on the last few decades.

As this AI hype cycle ramps up, we're actually simultaneously in the down ramp of _another_ AI hype cycle; the 5% for self-driving cars is going _very slowly indeed_, and people seem to have largely accepted that, while still predicting that the 5% for generative language models will be easy. It's odd.

(Though, also, I'm not convinced that it _is_ just a case of making a better ChatGPT; you could argue that if you want correct results, a generative language model just isn't the way to go at all, and that the future of these things mostly lies in being more convincingly wrong...)

How will it do that?

One of major problems of modern computer-based work is that there are too many people already in those roles, doing work that isn't needed. Case in point: the culling of tens of thousands of software engineers, people who would consider themselves to be doing 'bullshit jobs'.

It seems to me that the tendency to be confidently wrong is entirely baked into intelligence of all kinds. In terms of actual philosophical rationality, human reasoning is also much closer to cargo cults than to cogito ergo sum, and I think we're better for it.

I cannot but think that this approach of "Strong Opinions, Weakly Held" is a much stronger path forward towards AGI than what we had before.

The problem is that ChatGPT is about as useful as all the other dilettantes claiming to be polymaths. Shallow, unreliable knowledge on lots of things only gets you so far. Might be impressive at parties, but once there's real, hard work to do, these things fall apart.
Even if ChatGPT could only make us 10% better at solving the "easy" things but on a global scale, that is already a colossal benefit to society.
> I suppose it makes sense though. Denial is the default response when we face threats to our identity and sense of self worth.

Respectfully, that reads as needlessly combative within the context. It sounds like the blockchain proponents who say that the only people who are against cryptocurrencies are the ones who are “bitter for having missed the boat”.¹

It is possible and perfectly reasonable to identify problems in ChatGPT and similar technologies without feeling threatened. Simple example: someone who is retired and monetarily well off, whose way of living and sense of self worth are in no way affected by developments in AI, can still be critical and express valid concerns when these models tell you that it’s safe to boil a baby² or give other confident but absurdly wrong answers to important questions.

¹ I’m not saying that’s your intention, but consider that type of rhetoric may be counterproductive if you’re trying to make another understand your point of view.

² I passed by that specific example on Mastodon but I’m not finding it now.

I'm sorry I don't want it to get much smarter.

It you ask it to go through and comment code it does a pretty good job of that.

some things better than others(not that great at CSS)

need a basic definition of something. got it.

tell it to write a function it's not bad.

As a BA just tell it what your trying to do and what questions it should ask users. It will get some good ideas for you.

Want it to be a PM have create a loop asking every 10 minutes if your done yet.

Is it a senior engineer? no. can it pass a senior engineering interview? quite possibly.

debug code hit or miss.

I think the big thing it's not that great at front end code. It can't see so that probably makes sense. a fine-tuned version of clip that interacted with a browser would probably be pretty scary.

What's the point of letting it comment code? The programmer who reads the code can run it as well.
> The "I" in AI is just complete bullshit and I can't understand why so many people are in a awe

I agree.

And the worst thing is that the bullshit hype comes round every decade or so, and people run around like headless chickens insisting that "this time its different", and "this time its the REAL THING".

As you say, first(ish) there was ELIZA. Than this that and everything else. Then Autonomy and all that dot-com era jazz. Now with compute becoming more powerful and more compact, any man and his dog can stuff some AI bullshit where it doesn't belong.

I have seen comments below on this thread where people talk about "well, it's closing the gap". The thing you have to understand is that the gap will always exist. Ultimately you will always be asking a computer to do something. And computers are dumb. They are and will always be beholden to the humans that program them and the information that you feed them. The human will always have the upper hand at any tasks that require actual intelligence (i.e. thoughtful reasoning, adapting to rapidly changing events etc.).

This is not always true, see Chess.
AlphaGo as well. A few years back people were saying AI could never come close to beating a human at Go.
Man, if this were 1800 you'd be stating that man would never fly and the horse would never be supplanted by the engine. I honestly don't believe you have any scientific or rational reasoning for the point you are attempting to make in your post, because if you were you'd be stating that animal intelligence is magical.
> Man, if this were 1800 you'd be stating that man would never fly and the horse would never be supplanted by the engine.

I'm sorry, what sort of bullshit argument is that ?

Flight and engines are both natural evolution using natural physics and mechanics.

Artificial Intelligence is nothing but a square-peg-round-hole, when you have a sledgehammer everything looks like a nut scenario.

They are natural to you maybe with hindsight? Powered flight was most definitely not considered natural at the time. In fact, most attempts at flight were trying to mimic birds at first.
Flight and engines are natural evolution but intelligence is magic? Nature accomplished intelligence via random walk and it is a complicated mess because of it. To think that we cannot accomplish at least parts of intelligence is insane to me.
> And the worst thing is that the bullshit hype comes round every decade or so, and people run around like headless chickens insisting that "this time its different", and "this time its the REAL THING".

This. To answer the OPs question, this is what I'm fatigued about.

I'm glad we're making progress. It's a hell of a parlor trick. But the hype around it is astounding considering how often it's answers are completely wrong. People think computers are magic boxes, and so we must be just a few lever pulls away from making it correct all the time.

Or maybe my problem is that I've overestimated the average human's intelligence. If you can't tell ChatGPT apart from a good con-man, can we consider the Turing test passed? It's likely time for a redefinition of the Turing test.

Instead of AI making machines smarter, it seems that computers are making humans dumber. Perhaps the AI revolution is about dropping the level of average human intelligence to match the level of a computer. A mental race to the bottom?

I'm reminded of the old Rod Serling quote: We're developing a new citizenry. One that will be very selective about cereals and automobiles, but won't be able to think.

I'm having a really hard time following your argument. But absolutely agree we need to redefine the Turing test. Only problem is that I can no longer come up with any reasonable time-limited cognitive task that next year's AI would fail at, but a "typical human" would pass.
"Intelligence" is probably too nebulous a term for what it is we're trying to build. Like "pornography", its hard to rigidly define, but you know it when you see it.

I think "human level intelligence" is an emergent phenomenon arising from a variety of smaller cognitive subsystems working together to solve a problem. It does seem that ChatGPT and similar models have at least partially automated one of the subsystems in this model. Still, it can't reason, doesn't know it's wrong, and can't lie because it doesn't understand what a lie is. So it has a long way to go. But it's still real progress in the sense that it's allowing us to better see the dividing lines between the subsystems that make up general intelligence.

I think that we'll need to build a better systems level model of what general intelligence is and the pieces it's built out of. With a better defined model, we can come up with better tests for each subsystem. These tests will replace the Turing test.

>>Instead of AI making machines smarter, it seems that computers are making humans dumber. Perhaps the AI revolution is about dropping the level of average human intelligence to match the level of a computer. A mental race to the bottom?

I came here to make this comment. Thank you for doing it for me.

I remember feeling shocked when this article appeared in the Atlantic in 2008, "Is Google Making Us Stupid?": https://www.theatlantic.com/magazine/archive/2008/07/is-goog...

The existence of the article broke Betteridge's law for me. The fact that this phenomenon it is not more widely discussed describes the limit of human intelligence. Which brings me back around to the other side... perhaps we were never as intelligent as we suspected?

> perhaps we were never as intelligent as we suspected?

Yeah, I think you're right. Intelligence is just something our species has evolved as a strategy for survival. It isn't about intelligence, it's about survival.

The cognitive skills needed to survive/navigate/thrive in the digital era are very different than the cognitive skills required to survive in the pre-digital era.

We're biologically programmed through millions of years of evolution to survive in a world of scarcity. Intelligence used to be about tying together small bits of scarce information to find larger patterns so that we can better predict outcomes.

Those skills are being rendered more and more irrelevant in a world of information abundance. Perhaps the "best fit" humans of the future are those that possess new form of "intelligence", relying less on reason and more on the ability to quickly digest the firehose of data thrown at them 24-7.

If so, then the AI we were trying to build in the 1950s would necessarily be different than the AI that our grandchilden would find helpful.

You're dead on. Isn't it wild that despite our seemingly impressive intelligence, such insights never seem to rise to the level of... second nature.

I forgot to add something to my original post. >>"I remember feeling shocked when this article appeared in the Atlantic in 2008..."

At the time I was shocked that the question was even being asked!

I wonder if we will look back at this comment (and others like it) as similar to the infamous “takedown” of Dropbox when it was first posted on HN.

Time will tell, I certainly can’t predict.

I agree. I didn't understand the big deal that it passed a google interview either. IMO, that said more about the uselessness of the interview than the 'AI'.

Co-pilot has been semi-useful. It's faster than search SO, but like you said, I still have to review all the code and it's often wrong in subtle ways.

This is the meat of the issue - ChatGPT is exposing certain things a susceptible to bullshit attacks; humans have just been relatively bad at those.

It will turn out to be a useful tool for those who know what they’re asking about so they can check the answer quickly; but it will be USED by tons of people who don’t have a way of verifying the answers given.

I don't really think of ChatGPT as AI at this point, just an incredibly useful tool.
(comment deleted)
Like I said some days ago, I really wished that the hype would die or dwindle a bit. I'm working on my own AI side-projects, but the amount of BS and misinformation being put out everyday by new "AI experts" is fatiguing, yes.
The hype will die down fairly quickly. But this technology is obviously a huge deal. We've found a practical algorithm to turn more powerful hardware into better results. And hardware was still ramping up at an astonishing rate last time I checked.

It seems more likely that we'll surpass the hype than not in the next few decades. I think people have forgotten how quickly technology can move after the last 20 years of relative stability where more powerful hardware didn't really change what a computer can do.

ChatGPT is, of course, a great piece of software, but the huge hype is probably what it will be best remembered for. Also, since currently, AI is the exclusive playground of big corporations, to me it's a bit puzzling how some people can get so excited (and maintain that excitement) over something that they cannot control and have little hope of building it by themselves. I guess some are just more in love with technology, than with other things in life. Because, as everyone is probably well aware by now, more technology is the solution to every problem that ever faced mankind and will finally fix everything. :)
I assume that there will be an open and freely available model as large as ChatGPT within a year or so. Training costs are prohibitive but what about NSF grants?
I don't know about the NSF, or when will governments get in on AI, but you're probably right, the technology will become open source in a while, as it has happened in the past.

It looks much less likely for the cost of developing and training an AI system to come down for the time being, making it out of reach for most individuals.

When the PC revolution was happening, everyone interested had a good chance of getting in, they just needed some money to buy/rent a computer and learn to use it or program it.

Compared to that, the AI revolution doesn't seem to have the same quality.

The barrier to entry seems much much higher this time.

I do think that governments will have an interest in keeping around models that they're in control of, just like there is publicly funded boradcasting, you may want to be able to control all the biases of a widely-used model and not just import it from somewhere.