Not trying to discredit the premise but I do wonder how much this is from vertical scaling of model intelligence in weights vs the sheer amount of capital being thrown at training and horizontal scaling of agent swarms etc.
Anecdotally it seems like a lot of uplift continually comes from models doing more and exploring more of the search space, and we are being better at giving them embarrassingly parallel problems to Hoover up with infinite token spend.
I grok that you're not trying to discredit anything, or anything like that but what are you trying to do, if I may ask?
What information are you trying to either ask for or share with us?
The original article is pretty clear, agent swarms and token spend are AI automating AI, and the RSP attribution condition specifically excludes your capital thrown at training concern?
I guess my point is that the scaling is coming from capital explosion and it isn’t guaranteed that this will continue scaling as it did (the impending hardware and infrastructure crunch basically guarantees this).
The article is pretty strongly projecting this trend to continue but hasn’t acknowledged this factor.
This does rather remind me of how in the 80s we had a big boom in AI research, fueled in part by the development of new & expensive tools that everyone just had to have. And then it dissipated just as quickly when the companies who were pouring money into it started expecting to see a return on their investment.
Feels different now when the value of AI isn't just theoretical. Anthropic has like $80B ARR, and AI has really only transformed tech and a few limited spaces so far.
Not sure what work you do, but I don't really understand where such takes come from. As someone who works in software/ IT, this stuff is pure gold- it's multiplying not just the productivity but the abilities and intelligence of my entire team by a substantial factor. This has a very measurable value in money, and what the company pays for it now is just a small fraction of the value it gets, measured in terms of the salaries you'd need to pay to get the same level of performance. Anthropic could easily charge 5 times as much and still be convenient.
The technology people were developing in the 80s was also gold, in terms of its usefulness. Heck, much of it forms an essential part of the coding harnesses we all rely on nowadays.
But that didn’t stop the AI winter from coming. Because a technology being transformative does not protect it from over-investment. History perhaps suggests the opposite is true.
Nor did it prevent the AI companies of the 80s from going bankrupt when they got undercut by the commoditization that naturally came with the maturation of the technology.
>Feels different now when the value of AI isn't just theoretical.
We're coming up on four years since ChatGPT launched, and the only thing LLMs seem to excel at is the movement of imaginary little green pieces of paper.
I don't think AGI is inevitable from this point on but I don't understand this sentiment. There are a number of knowledge work sectors where LLMs are clearly here to stay and have truly fulfilled some of the core promises.
I don't think it'll continue on linearly, but it will likely continue.
We're getting the Claude efficiency talks at work. We're looking at all sort of optimizations, including local models and graphify, to reduce spend. There are people dedicating their nights and weekends to dropping token consumption and cache rebuilding. A few of these will likely catch on and be brought into common use. That will relieve at least a bit of the pressure, and enable some of the scaling to continue.
> the scaling is coming from capital explosion and it isn’t guaranteed that this will continue scaling
Not at all. Have you tried local AI lately? Qwen 3.8 is legitimately much better than SoTA from 1 year ago (Opus 4.1). You can run it on a consumer GPU.
Qwen Next Flash runs on a DGX spark well and is as good as some things you need datacenter class GPU's to run even today.
Check out Artificial Analysis if you haven't. Sometimes you don't realize how fast these things are moving, but to compare things from a year ago to today is just wild.
They aren't just getting better, they're getting more efficient by many orders of magnitude.
1. Intelligence per dollar spent/token output seems to be increasing rapidly. Opus 4 was $15/75 per million input/output tokens when released. Opus 5.5, which is significantly better, is $4/15. Sure it cost more money upfront to train Opus 5.5, or the bigger model (Fable 5.5) which tuned it, but once it is reached, there are massive diffusive effects.
2. The capital equation is still favoring more spend. Anthropic's revenue is growing rapidly and they are profitable on inference. Huge upfront spend for realized gains later on is standard for R&D. As long as serving the models is profitable and generates more value than the energy/compute consumed
3. The search space is too large to assume that improvement in AI assisted research is just models dumbly trying every combination until something works. For every additional bit searched, the space doubles, so even a trivially small combinatorial exploration incurs prohibitively large costs. Intelligent searching which continuously prunes that space along dimensions of effectiveness is not a trivial skill and is among the emergent accoutrements of these smarter models.
> 2. The capital equation is still favoring more spend. Anthropic's revenue is growing rapidly and they are profitable on inference. Huge upfront spend for realized gains later on is standard for R&D. As long as serving the models is profitable and generates more value than the energy/compute consumed
But they are essentially fighting for smaller and smaller part of the pie. The better open models are the less you need frontier models, so the frontier AI companies are essentially fighting for market that will shrink the better AI is.
Sure there is inertia and the usual corporate enterprise nonsense of "we only buy stuff from other big companies regardless the cost" that keeps the licenses going but we're quickly getting into point where you don't need any of the frontier models to do majority of the tasks
I don't fully get this line of argument. I am a massive advocate for open source... everything basically, but especially in ML. And open models are making incredible strides. But I'm not sure that models getting good enough for most tasks, making the absolute frontier models (Astra/Fable class) unnecessary for most people, puts the frontier labs in a fundamentally disadvantaged position.
I mainly think this because frontier labs already try to compete at this tier. For example, 6.1 Sol has identical pricing to Kimi K3.1, but with a lower price per task due to Kimi's verbosity. It seems to score roughly the same on most benchmarks, with some (like Artificial Analysis) putting 6.1 Sol significantly higher. If you need to step down, 6.1 Terra is also cheaper than K3.1 or K3 and performs well relative to their benchmarks. Opus 5.5 and much of Anthropic's recent push seems to confirm this as a major area of focus as well.
I still use a lot of open models and harnesses, but I don't know if cost competitiveness (at least not with the current crop of open models) will ultimately drive a huge chunk of average users away.
Things could change, but fundamentally, large labs aren't precluded from training more cost effective models if there is significant demand for them, which is what we're seeing now.
Did I ignore them? I said that I love open models. If it came across like I'm putting down open models, then I'm genuinely sorry--I love open source ML, have been a involved in it for a long time, and think it's incredibly important. It's also just fun to work with open models, beyond the benefits to privacy and flexibility.
I'm just talking about the idea that frontier labs will lose market share to providers that sell open model inference at sub-Fable prices. And I'm not convinced that this specifically is a threat to frontier labs, because they also train and sell models that fit that demand.
6.1 Sol is not a “frontier model” and does not command the pricing of one. The reason it exists is because of competition from non-frontier models that are “good enough”.
In fact, you kind of made the same argument. As “good enough” models commoditized, the use cases requiring frontier models and frontier model pricing will shrink.
And, even if performance is not as good, many will prefer open models.
I think I did address the argument? This being the key quote:
"The better open models are the less you need frontier models, so the frontier AI companies are essentially fighting for market that will shrink the better AI is."
The implication in this is that frontier labs' market is exclusively people who need the absolute frontier, i.e. Astra or Fable. And the point I made was that frontier labs are clearly prioritizing the broader market of people who want more cost effective models, and that if OpenAI/Anthropic offers models that are price competitive with the top open models (before factoring in subscription discounts and corporate lock-in), then I don't see how that leads to the average user walking away from the labs in favor of self-hosting or relying OpenRouter/Together/etc.
I also went out of my way to clarify that I personally prefer and use open models wherever possible. I'm not saying that there isn't a real group of people who prefer them. I'm specifically saying that frontier labs also sell "good enough" models and that there's nothing preventing them from training models to fit the demands of users. It's not as if training a frontier model boxes them in such that they can't release an Opus 5.5 or a 6.1 Terra.
> if OpenAI/Anthropic offers models that are price competitive with the top open models (before factoring in subscription discounts and corporate lock-in), then I don't see how that leads to the average user walking away from the labs in favor of self-hosting or relying OpenRouter/Together/etc.
The premise is this: frontier models seem to advance, and OSS models also seem to advance in terms of cost efficiency etc. It's like a race with say a 6 month (WLOG) gap between frontier and OSS models.
For frontier models to win perpetually against OSS, they will need to have a perpetual 6-month lead, or develop an edge OSS can't replicate for some architectural reasons (say MacOS vs Linux). If say a frontier lab trips because of a bad quarter or some approach temporarily maxxing out, they may not be able to maintain an edge vs OSS any more. It is a structural argument, not an empirical description of the advantages today.
What is the unit of measure for “intelligence per dollar”, and how do you measure it? Without a concrete definition, that first assertion is not even wrong.
I would also like to point out that “profitable on inference” is similarly meaningless, because it’s based on a figure that arbitrarily discards whatever bits of their operating costs they need to to get to the claim that they’re profitable on inference. A lot of these numbers Anthropic has been spewing mostly exist to distract people who don’t understand why GAAP exists in the first place.
>What is the unit of measure for “intelligence per dollar”, and how do you measure it?
It's going to depend on the task (and this is prone to cheating), but you can get a decent idea by looking at task benchmarks and the cost of completing them.
I was inclined to agree but Anthropic's latest S-1 statement only includes a paper loss due to their value going up. They're objectively making money now, the only question is can they keep it up and can they grow enough to be able to pay the bills they've signed up for in future?
But benchmark performance is not the same thing as intelligence.
Which is a really important distinction. One measures specific points on the jagged boundary. The other is a word that blithely paints it all with a single brush, and sweeps the jagged boundary problem under the rug in the process.
It really is important to keep track of that. Whenever we talk about “intelligence” as if it were a well-defined (let alone quantifiable) thing, we fall into the kind of wooly thinking that lets anyone say pretty much anything. We abandon the realm of grounded statements in favor of pure vibes.
Eh, I get what you mean. AI especially seems to have performance that is much more difficult to reason about than human intelligence (where most aspects of mental capability are surprisingly strongly correlated). But in this context, if you broadly focus on intelligence as the ability to complete problem-solving tasks, models have been improving pretty uniformly at the same time as getting cheaper.
> Intelligence per dollar spent/token output seems to be increasing rapidly.
I would argue this is more due to market economics than actual inherent cost. Opus 4 was when the open models started catching up. The R1 bootstrapping chain of thought paper via RL paper was released a couple of months before Opus came out and the open models started to significantly catch up after that. The duopoly was effectively broken and they lost their price setting power.
> Anthropic has made roughly two years of historical progress over the last year [...] We are confident that the acceleration was not due to an acceleration in the scaling of training compute.
Note that Anthropic has essentially heavily censored this report, and the authors themselves say "the public has to rely on METR’s judgment much more than would be ideal".
"The industry is not prepared to handle this safely"
After the election, titles like this just irritate me. Every piece written about AI progress is:
"The machines taking jobs, flooding the internet with sludge and poisoning our online experiences, are getting better even faster! Isn't this incredible? Yeah but sorry we also think it's only riskier than ever to deploy these systems at scale. That being said, we're still going to do it, go public and take institutional money from people's retirement funds, destroy the job market, and make 3000 Anthropic employees and 500 investors really rich. There wasn't anything we could do besides write a lot of ink on it and be shocked we lost every local data center referendum."
The author of this article works for METR and not Anthropic, their whole job is supposed to be providing independent oversight on the labs. They are definitely not in the "this is incredible" crowd, they are the ones asking for development and deployment to slow down.
there's considerable overlap between the "we need to slow down" crowd and the "this is amazing crowd". The Yudkowsky/rationalist crowd are still Utopianists, they just think we need to be careful how we achieve it.
Antrophic shouldn't self-regulate, it's government responsibility to do that. Unfortunately US has demented manchild as a president so that doesn't work like it should.
The very same report also found that AI could not be deployed outside of a narrow domain of verifiable rewards … so take it with grains, or mountains of salt. It’s also extremely unclear how the productivity measures are calculated. They surely don’t seem to translate into economic output. Antrophic has seen its valuation corrected by 50% and possibly more… so yeah. This all sounds very fishy.
Yes, it is the local model narrative that BigAI likes because it will not happen but people are deluded into thinking that there is no danger of an oligarchy.
Mathematicians for example have used it as cope for their current reliance on unsavory corporations, telling themselves "just one more year, and I can solve a millennium problem on my Asus Strix".
As long as we believe, we can imagine it is true. All seeing Oracle: how do you command me? Praise his name, bringer of many miracles - he who definitely can't be manipulated from the backend to provide biased information that suites a person's malicious purpose. Ya totally.
i'm also wondering whether there are confounding factors like more domain experts and talents joining the AI research field because there are more interesting problems to be solved (or other 2nd order effects of AI models being more powerful).
Another data point: look at the number of submissions to Neurips, ICLR, etc. Now, I agree that a submission does not by itself imply groundbreaking resesarch, but the fact that people are able to try out/vet ideas much faster now makes a difference.
In my own work, I've found that it's much faster to brainstorm with a frontier model like Opus. I'll have an idea and be able to get a working implementation and try it out on toy problems within O(minutes), instead of O(weeks) earlier.
Trying out problems and making tangible progress towards AGI are entirely different categories. Likewise what’s changed in terms of paper submissions is also the amount of slop. Most reviewers say that they haven’t seen a lot more papers that are good but rather they’ve seen a lot more papers and almost all of them are terrible slop that’s desk rejected. This is happening in OSS too where a lot of projects are closing the gate at the onslaught of terrible contributions.
I agree in general that AI accelerates AI research, but good lord neurips iclr papers (and reviews might I add...) are the worst example of that. It's all complete slop that is totally meaningless.
A lot of great papers by great people had a lot of AI involved and accelerating it of course, but the majority is meaningless (sometimes autoresearch) slop.
In October/September my LinkedIn wall was probably… 55% people complaining that they were getting garbage AI submissions to the conferences they review for, and 35% people complaining that they got garbage AI reviews back from their conference submission… Of course, I have no idea if that means anything given how aggressively that site bubbles your feed.
Yeah it already helps me too with finetuning small models for image stuff.
Im not telling it what to do, i ask it for ar eport/analysis and research and read its article and tell it what to implement based on 'feelings' and it works.
Limit is hardware and tokens and an AI Researcher not having hardware or tokens is probably not what they refer to.
Alone normal researchers were quite handicaped: converting data from tool a to python library x, understanding how tools work, doing basic google research etc.
I find this as well for models that make it to production.
On a more fun note, I also find just trying some hair brained idea is now extremely cheap - so why not? I used to have to weigh up spending a week of effort on something likely not to work and now I can now get spit out test in 20 minutes.
I do agree. AI isn't "free" the way silicon valley products are.
Therefore it cannot avoid taxes the way they did before (using shell companies to route the money through tax havens). It'll be expensive, every government will double the price or more, and you're selling to consumers, and the products will be at the mercy of every government where they try to sell AI and other services.
People don't get how serious this is. If AI doubles revenue but drives silicon valley firms' tax rates from 5% to 10%, it's a loss.
Add to that that it's something silicon valley firms have never done before.
Does anyone remember the book "The Singularity is Near?"
I read it about 20 years ago, and once it got into nanobots and Kurzweil's vision that the whole universe would be intelligent, I didn't find the book credible.
But I'm wondering if the short-term prediction, that in the 2030s we won't be able to anticipate what tech does for us, (IE, the Singularity,) is useful?
I've read most of Kurzweil's books and been a believer in the Moore's law like stuff but always skeptical on the nanobots. He takes ideas from a variety of sources and the nanobot idea seems to be quite questionable.
In the original article he noted that getting to a 2x increase in capabilities would be crazy and require ~40x researcher productivity; that’s exactly the increase the new data now shows, and thus a huge number of the assumptions made in this earlier post about how it’s unlikely we would get to this point have literally already been proven wrong less than two months later.
I expect to see an update from Ramez moving substantially more in the direction of RSI in fact very much increasing the odds of reaching ASI soon.
I’m very curious what the major bottleneck is in the current landscape. I can only think of three possibilities: 1) model capabilities, 2) human researchers, or 3) available computers resources.
Given that frontier models are leapfrogging and massive infrastructure build is ongoing, I don’t see why any of these three couldn’t be overcome.
Interesting write-up, but I feel like a lot of its argument relies on the interpretation of ECI scores as merely showing linear improvement in AI capabilities. Sure, Epoch AI themselves claim that ECI's scale is linear, meaning that a large increase in capabilities should result in a large increase in ECI. Yet, the widely perceived jump in coding abilities in late 2025 that started making agentic coding actually productive doesn't really show on the software engineering ECI.
The ECI is basically an aggregate of benchmark scores, but benchmark complexity might not increase linearly. In fact, if LLM capabilities grew exponentially, then new benchmarks would also need to increase their difficulty / complexity exponentially to stay unsaturated and relevant.
I suspect that if models were able to be aware of and answer questions about their internal vectors, they would be much smarter and able to progress faster.
AI is accelerating my work...work which only exists because of AI.
If I had to do my individual tasks without AI, I would be slower. But if I take a step back, I'm really not changing the world any faster than I was before.
Was yours better because it used my phrasing against me?
Anyway, I'm surprised to see it flagged. I think it's the logical conclusion, that if you are having your work accelerated, and you also aren't changing the world any faster, then your work is not focused on changing the world--it needs to be more "important."
If it was world-changing already, then doing more of it would, in fact, accelerate changing the world.
But they're here saying something different, blaming the AI instead of their own focus.
For all my user facing work I’m now required to do an AI summary of what I did and the impact. That summary is then graded by AI, against an AI generated rubric. For my yearly review, I use AI to summarize my summaries, and my managers AI summarizes the AI generated grades.
I’m also required to explain how I use AI in my work. I must plan all of my future work, with an AI tool. Another AI assesses whether my work is aligning with planned work.
Before I work and after I work, there is a million circular references that exist solely because of AI. These have scaled because it’s so easy to do with AI, but the we’re wasting so much time doing a ton of extra things very efficiently.
Let me do my job. My manager should know how I’m doing. Maybe highlight key wins. CSATs/etc go out every year.
So….you have outsourced multiple layers of middle-management to AI? I am not sure getting AI to write and review your TPS reports is going to move the needle.
The people who in theory read these are managers 3+ levels above me, so I honestly have no idea. I know my manager reads mine, but I haven’t seen any outcome from that.
Even before my last job had internally approved AI, my manager told me how he was using chatgpt on his personal computer to prepare our yearly reviews. Then when they released AI internally a few weeks later, it was still review time, and their first advertised use case was for doing your own self eval and the evals for your reports. I am willing to deal with vibed code but turning over employee evals to AI too is a sign of truly terrible management
Its true, a part of me died when an "AI first" company after adoption of AI introduced a mandatory design report (waterfall all over again) for EACH jira ticket before starting to code
Imagine developers move faster because of AI programming! No lets create more slop for other slopmachines
It’s true, design has made a big comeback. Now we actually have time to design things carefully and try the design easily with LLM. This was just too expensive before but now the calculation changed completely. I am now using spec graphs to design systems and I have been able to create things much more advanced than I was able to before by myself. This is huge! People are still figuring this out for themselves. But now we can design things, have multiple rounds of adjustments in an afternoon, have things implemented differently and actually measure the trade offs. The final product is so much better it’s crazy. If you had an eye for design before, now this is your world. If you didn’t, yeah your will probably just bias the LLM into producing shit solutions, even if it tries to warn you about it, it still follows orders in the end.
There's this joke like "I used to have 5 unfinished side projects, but AI changed everything. Now I have 30 unfinished side projects". And in my case it's not even a joke
The definition of "finished" has changed for me also.
More and more, I am using half finished as a polished product, but pragmatically very useful software I have built myself.
More importantly, I see no reason to polish this into a product. It is more useful in this half finished, easily moldable state.
Like it use to make sense to bake the clay and sell the hardened output but now I want to keep the soft clay so it can be molded to my changing needs. I keep adding to a collection of these moldable states.
I've definitely found that while "AI makes getting from 0-80 done much easier than the 80-100" holds true, it is much easier to get my work to a quality bar that I'm excited to share with my peers...
Well ai made it extremely easy to finish things. Projects are usually left unfinished because the last mile is filled with boring and tedious tasks we do not like to do. AI loves those tasks. I have finished everything I bothered to, most of the time I still wrote stuff by hand, the AI just fixed tests for me (usually OS specific stuff that I can’t test locally which is a pain), told me what I was doing wrong so it was trivial for me to fix it, and checked my documentation which invariably got out of date.
I mean, for many years I worked on software applications for provisioning and controlling cloud orchestration platforms that only existed... to serve software applications.
It's not that deep—there were many more verticals outside of control plans, and there are now many more verticals outside of machine learning.
what if the greatest threat is simply there will be hedonistic experiences so great most humans will find them irresistible to the point of shorter lives.
Could be from a sandwich that tastes like the best sandwich ever, but is super bad for your cholesterol or something.
I dont know where Im going with this.
Actually did some research I think the term is "Wireheading" from a story/theory from the 50s
This is already true? Ultra processed foods and tiktok are both deleterious and both wildly successful.
Maybe what you are concerned with is a theoretical sandwich that was unfailingly addictive so that no one could escape its effects.
Nukes don’t kill people, people with nukes kill people. How do you feel about giving everyone on earth access to nukes?
A simple thought experiment - imagine AI discovers a way to make nuclear bombs capable of destroying the entire world radically more cheaply and accessibly with common items anyone can purchase. AI can help anyone build these nukes trivially. Even if the AI and the nukes don’t kill people on their own, do you want literally every human on earth to have this power in their pocket? Do you think it would be a net positive outcome for us vs a world where that’s not the case?
Yeah, let's also ban 3d printers, drone motors, and YouTube chemistry videos. If you want to be dangerous, there's a world of opportunities already, with existing open models quite sufficient for a lot of interesting cases.
This seems kind of aggrandizing when BigAI shifts goalposts constantly whether it's "we created AGI" by a definition that an autopilot in a commercial plane fulfills or "Maths research" where everyone else has to do the actual research.
308 comments
[ 0.25 ms ] story [ 20.0 ms ] threadAnecdotally it seems like a lot of uplift continually comes from models doing more and exploring more of the search space, and we are being better at giving them embarrassingly parallel problems to Hoover up with infinite token spend.
What information are you trying to either ask for or share with us?
The original article is pretty clear, agent swarms and token spend are AI automating AI, and the RSP attribution condition specifically excludes your capital thrown at training concern?
But, then as now, what they didn’t have was profits.
But that didn’t stop the AI winter from coming. Because a technology being transformative does not protect it from over-investment. History perhaps suggests the opposite is true.
Nor did it prevent the AI companies of the 80s from going bankrupt when they got undercut by the commoditization that naturally came with the maturation of the technology.
We're coming up on four years since ChatGPT launched, and the only thing LLMs seem to excel at is the movement of imaginary little green pieces of paper.
We're getting the Claude efficiency talks at work. We're looking at all sort of optimizations, including local models and graphify, to reduce spend. There are people dedicating their nights and weekends to dropping token consumption and cache rebuilding. A few of these will likely catch on and be brought into common use. That will relieve at least a bit of the pressure, and enable some of the scaling to continue.
Not at all. Have you tried local AI lately? Qwen 3.8 is legitimately much better than SoTA from 1 year ago (Opus 4.1). You can run it on a consumer GPU.
Qwen Next Flash runs on a DGX spark well and is as good as some things you need datacenter class GPU's to run even today.
Check out Artificial Analysis if you haven't. Sometimes you don't realize how fast these things are moving, but to compare things from a year ago to today is just wild.
They aren't just getting better, they're getting more efficient by many orders of magnitude.
https://artificialanalysis.ai/#intelligence
the same thing you are:
> if I may ask?
asking.
1. Intelligence per dollar spent/token output seems to be increasing rapidly. Opus 4 was $15/75 per million input/output tokens when released. Opus 5.5, which is significantly better, is $4/15. Sure it cost more money upfront to train Opus 5.5, or the bigger model (Fable 5.5) which tuned it, but once it is reached, there are massive diffusive effects.
2. The capital equation is still favoring more spend. Anthropic's revenue is growing rapidly and they are profitable on inference. Huge upfront spend for realized gains later on is standard for R&D. As long as serving the models is profitable and generates more value than the energy/compute consumed
3. The search space is too large to assume that improvement in AI assisted research is just models dumbly trying every combination until something works. For every additional bit searched, the space doubles, so even a trivially small combinatorial exploration incurs prohibitively large costs. Intelligent searching which continuously prunes that space along dimensions of effectiveness is not a trivial skill and is among the emergent accoutrements of these smarter models.
But they are essentially fighting for smaller and smaller part of the pie. The better open models are the less you need frontier models, so the frontier AI companies are essentially fighting for market that will shrink the better AI is.
Sure there is inertia and the usual corporate enterprise nonsense of "we only buy stuff from other big companies regardless the cost" that keeps the licenses going but we're quickly getting into point where you don't need any of the frontier models to do majority of the tasks
I mainly think this because frontier labs already try to compete at this tier. For example, 6.1 Sol has identical pricing to Kimi K3.1, but with a lower price per task due to Kimi's verbosity. It seems to score roughly the same on most benchmarks, with some (like Artificial Analysis) putting 6.1 Sol significantly higher. If you need to step down, 6.1 Terra is also cheaper than K3.1 or K3 and performs well relative to their benchmarks. Opus 5.5 and much of Anthropic's recent push seems to confirm this as a major area of focus as well.
I still use a lot of open models and harnesses, but I don't know if cost competitiveness (at least not with the current crop of open models) will ultimately drive a huge chunk of average users away.
Things could change, but fundamentally, large labs aren't precluded from training more cost effective models if there is significant demand for them, which is what we're seeing now.
I'm just talking about the idea that frontier labs will lose market share to providers that sell open model inference at sub-Fable prices. And I'm not convinced that this specifically is a threat to frontier labs, because they also train and sell models that fit that demand.
6.1 Sol is not a “frontier model” and does not command the pricing of one. The reason it exists is because of competition from non-frontier models that are “good enough”.
In fact, you kind of made the same argument. As “good enough” models commoditized, the use cases requiring frontier models and frontier model pricing will shrink.
And, even if performance is not as good, many will prefer open models.
"The better open models are the less you need frontier models, so the frontier AI companies are essentially fighting for market that will shrink the better AI is."
The implication in this is that frontier labs' market is exclusively people who need the absolute frontier, i.e. Astra or Fable. And the point I made was that frontier labs are clearly prioritizing the broader market of people who want more cost effective models, and that if OpenAI/Anthropic offers models that are price competitive with the top open models (before factoring in subscription discounts and corporate lock-in), then I don't see how that leads to the average user walking away from the labs in favor of self-hosting or relying OpenRouter/Together/etc.
I also went out of my way to clarify that I personally prefer and use open models wherever possible. I'm not saying that there isn't a real group of people who prefer them. I'm specifically saying that frontier labs also sell "good enough" models and that there's nothing preventing them from training models to fit the demands of users. It's not as if training a frontier model boxes them in such that they can't release an Opus 5.5 or a 6.1 Terra.
The premise is this: frontier models seem to advance, and OSS models also seem to advance in terms of cost efficiency etc. It's like a race with say a 6 month (WLOG) gap between frontier and OSS models.
For frontier models to win perpetually against OSS, they will need to have a perpetual 6-month lead, or develop an edge OSS can't replicate for some architectural reasons (say MacOS vs Linux). If say a frontier lab trips because of a bad quarter or some approach temporarily maxxing out, they may not be able to maintain an edge vs OSS any more. It is a structural argument, not an empirical description of the advantages today.
I would also like to point out that “profitable on inference” is similarly meaningless, because it’s based on a figure that arbitrarily discards whatever bits of their operating costs they need to to get to the claim that they’re profitable on inference. A lot of these numbers Anthropic has been spewing mostly exist to distract people who don’t understand why GAAP exists in the first place.
It's going to depend on the task (and this is prone to cheating), but you can get a decent idea by looking at task benchmarks and the cost of completing them.
I was inclined to agree but Anthropic's latest S-1 statement only includes a paper loss due to their value going up. They're objectively making money now, the only question is can they keep it up and can they grow enough to be able to pay the bills they've signed up for in future?
Which is a really important distinction. One measures specific points on the jagged boundary. The other is a word that blithely paints it all with a single brush, and sweeps the jagged boundary problem under the rug in the process.
It really is important to keep track of that. Whenever we talk about “intelligence” as if it were a well-defined (let alone quantifiable) thing, we fall into the kind of wooly thinking that lets anyone say pretty much anything. We abandon the realm of grounded statements in favor of pure vibes.
I would argue this is more due to market economics than actual inherent cost. Opus 4 was when the open models started catching up. The R1 bootstrapping chain of thought paper via RL paper was released a couple of months before Opus came out and the open models started to significantly catch up after that. The duopoly was effectively broken and they lost their price setting power.
> Anthropic has made roughly two years of historical progress over the last year [...] We are confident that the acceleration was not due to an acceleration in the scaling of training compute.
Note that Anthropic has essentially heavily censored this report, and the authors themselves say "the public has to rely on METR’s judgment much more than would be ideal".
Perhaps scaling to fable for the first time also gave Opus a push (distillation?)
After the election, titles like this just irritate me. Every piece written about AI progress is:
"The machines taking jobs, flooding the internet with sludge and poisoning our online experiences, are getting better even faster! Isn't this incredible? Yeah but sorry we also think it's only riskier than ever to deploy these systems at scale. That being said, we're still going to do it, go public and take institutional money from people's retirement funds, destroy the job market, and make 3000 Anthropic employees and 500 investors really rich. There wasn't anything we could do besides write a lot of ink on it and be shocked we lost every local data center referendum."
The author of this article works for METR and not Anthropic, their whole job is supposed to be providing independent oversight on the labs. They are definitely not in the "this is incredible" crowd, they are the ones asking for development and deployment to slow down.
https://metr.org/assets/images/ai-rd-acceleration-assessment...
Mathematicians for example have used it as cope for their current reliance on unsavory corporations, telling themselves "just one more year, and I can solve a millennium problem on my Asus Strix".
In my own work, I've found that it's much faster to brainstorm with a frontier model like Opus. I'll have an idea and be able to get a working implementation and try it out on toy problems within O(minutes), instead of O(weeks) earlier.
They seem very strongly related to me.
A lot of great papers by great people had a lot of AI involved and accelerating it of course, but the majority is meaningless (sometimes autoresearch) slop.
Specifically talking about recent neurips/iclr.
Im not telling it what to do, i ask it for ar eport/analysis and research and read its article and tell it what to implement based on 'feelings' and it works.
Limit is hardware and tokens and an AI Researcher not having hardware or tokens is probably not what they refer to.
Alone normal researchers were quite handicaped: converting data from tool a to python library x, understanding how tools work, doing basic google research etc.
On a more fun note, I also find just trying some hair brained idea is now extremely cheap - so why not? I used to have to weigh up spending a week of effort on something likely not to work and now I can now get spit out test in 20 minutes.
I always wanted to build a specific event/sequence in one LSD trip into a small game to give that experience some kind of playability.
Then I saw some 4k demo, wanted to see how good it is in doing this.
Its really fun yeah :)
Therefore it cannot avoid taxes the way they did before (using shell companies to route the money through tax havens). It'll be expensive, every government will double the price or more, and you're selling to consumers, and the products will be at the mercy of every government where they try to sell AI and other services.
People don't get how serious this is. If AI doubles revenue but drives silicon valley firms' tax rates from 5% to 10%, it's a loss.
Add to that that it's something silicon valley firms have never done before.
Would you rather make $1,000,000 and pay $50,000 in taxes, or make $2,000,000 and pay $200,000 in taxes?
The choice is obvious
I read it about 20 years ago, and once it got into nanobots and Kurzweil's vision that the whole universe would be intelligent, I didn't find the book credible.
But I'm wondering if the short-term prediction, that in the 2030s we won't be able to anticipate what tech does for us, (IE, the Singularity,) is useful?
https://en.wikipedia.org/wiki/The_Singularity_Is_Near
Even if he turns out to be a decade or two off, that’s basically exactly correct in the sweep of history sense.
Wild time to be alive.
That is factually incorrect, and seems to be much of the premise behind the authors way of thinking.
In the original article he noted that getting to a 2x increase in capabilities would be crazy and require ~40x researcher productivity; that’s exactly the increase the new data now shows, and thus a huge number of the assumptions made in this earlier post about how it’s unlikely we would get to this point have literally already been proven wrong less than two months later.
I expect to see an update from Ramez moving substantially more in the direction of RSI in fact very much increasing the odds of reaching ASI soon.
Given that frontier models are leapfrogging and massive infrastructure build is ongoing, I don’t see why any of these three couldn’t be overcome.
The ECI is basically an aggregate of benchmark scores, but benchmark complexity might not increase linearly. In fact, if LLM capabilities grew exponentially, then new benchmarks would also need to increase their difficulty / complexity exponentially to stay unsaturated and relevant.
Opus 5.5 is so amazing I'm not surprised if it can improve itself.
If I had to do my individual tasks without AI, I would be slower. But if I take a step back, I'm really not changing the world any faster than I was before.
Do more important work.
Anyway, I'm surprised to see it flagged. I think it's the logical conclusion, that if you are having your work accelerated, and you also aren't changing the world any faster, then your work is not focused on changing the world--it needs to be more "important."
If it was world-changing already, then doing more of it would, in fact, accelerate changing the world.
But they're here saying something different, blaming the AI instead of their own focus.
Could you elaborate? What work are you doing which only exists because of AI?
I’m also required to explain how I use AI in my work. I must plan all of my future work, with an AI tool. Another AI assesses whether my work is aligning with planned work.
Before I work and after I work, there is a million circular references that exist solely because of AI. These have scaled because it’s so easy to do with AI, but the we’re wasting so much time doing a ton of extra things very efficiently.
Let me do my job. My manager should know how I’m doing. Maybe highlight key wins. CSATs/etc go out every year.
This sounds truly Kafkaesque. And thank you for finally allowing me to use that word in earnest.
Kafka and Byzantium had a baby.
Imagine developers move faster because of AI programming! No lets create more slop for other slopmachines
More and more, I am using half finished as a polished product, but pragmatically very useful software I have built myself.
More importantly, I see no reason to polish this into a product. It is more useful in this half finished, easily moldable state.
Like it use to make sense to bake the clay and sell the hardened output but now I want to keep the soft clay so it can be molded to my changing needs. I keep adding to a collection of these moldable states.
But its a lot less effort that when we didn't have LLM's.
Interestingly it does seem that it's still enough effort to prevent most folks from making that step though.
It's not that deep—there were many more verticals outside of control plans, and there are now many more verticals outside of machine learning.
Or am I missing the point?
Could be from a sandwich that tastes like the best sandwich ever, but is super bad for your cholesterol or something.
I dont know where Im going with this.
Actually did some research I think the term is "Wireheading" from a story/theory from the 50s
(Infinite Jest)
https://en.wikipedia.org/wiki/Wirehead_(science_fiction)
AIs do not kill people, people kill people.
A simple thought experiment - imagine AI discovers a way to make nuclear bombs capable of destroying the entire world radically more cheaply and accessibly with common items anyone can purchase. AI can help anyone build these nukes trivially. Even if the AI and the nukes don’t kill people on their own, do you want literally every human on earth to have this power in their pocket? Do you think it would be a net positive outcome for us vs a world where that’s not the case?
https://nickbostrom.com/papers/vulnerable.pdf
Shifting goalposts is shifting goalposts, doesn't matter if it's a hackernews comment or a 'BigAI' marketing blurb.
My statement was that AI is dramatically accelerating research that has objective goals, idgaf what 'BigAI' does.