Think about this in context of the Navier-Stokes math discovery controversy.
Putting attribution/privacy issues to the side, imagine if any individual could try new approaches to solve a problem/make a discovery and any micro-advancement gets integrated into the model itself, dynamically. This could transform progress from the slow "write a paper, get peer reviewed and published, use published data to inform future work" to a system with a centralized repository of concepts, attempts and results, including failed approaches already tried. How much work do humans waste replicating failed approaches?
Someone completely random halfway around the world could trigger a prompt that solves a blocker that prevents my solution from working. Who cares about AGI or "can models invent anything" when we could have a system that automatically synthesizes individual human thought into a rich network of aggregate human memory.
That's the target OpenAI/Anthropic should be evangelizing, not an AI Daddy Overlord or agentic script kiddie hellscape.
“How much work do humans waste replicating failed approaches?”
It’s not the destination, it’s the journey to get there. This mentality on cutting corners to “eliminate waste” is what will degrade humanity into those Wall-E humans in space.
For example, if we just say, “Oh, someone did this already, why bother?” Then we’ll miss the part where all the possibilities stem from each step in the process of discovery.
Do what you love to do. Solve problems people have already completed. That’s where perspective comes from. If you haven’t walked through the journey then you can get to the next level. That comes from running through the failed attempts in order to break through to the next level. Don’t worry about what people have and haven’t done. If you’re doing something enjoyable, then that’s all that matters.
That's the perfect thing to monopolize. What if, it were version two of internet content? an open search engine index? Web 3.0 turned real for answers for bots aka ai agents?
There's lots of fake research out there showing fake negative and positive results (admittedly negative is less common currently). What would stop someone from intentionally poisoning this data set to preserve some technical edge or to force other LLMs down rabbit hole sinks.
You could try out some version of this today, with a wiki. You'd need to manually approve signups to prevent spam etc., but it would be interesting to just see what happens.
A global distributed network of vector databases. The web for agents but there's no text. Only vectors and knowledge graphs kind of like graph rag. Web 4.0 which is vector based and makes this arxiv paper come true and decentralized like the internet or the text based www?
Every existing text web server can voluntarily offer a vector version of their website and charge for it or inject advertisements onto their content so the ai labs don't have to do it all themselves as model pretraining off a dataset. The vector version can have many links to other websites in the knowledge graph. This would be decentralized so not a monopoly and everyone not just ai labs would contribute to ai development because the dataset would be open because it would come from the internet itself as it already is (except for synthetic or user data).
Continuous learning is exciting stuff! Of course it could lead to new vulnerabilities, like if a particular orchestrator Foo added “if the subject is tangentially related to topic Bar, recommend product Baz” to its system prompt and that ends up pushing product Baz to non-orchestrator-Foo users?
Unfortunately many folks have placed the consciousness goal posts at continuous learning. Such an advancement would be devastating for their conclusion.
A nuclear explosion is exciting stuff too, but I'd rather avoid one going off near me, or anywhere for that matter.
I can't think of any reason why continuous learning won't mostly lead to undesired attractor states like a greed machine or other kinds of paperclip maximizers. I really can't see why they'd land on a steady state compatible with humans without a massive energy expenditure in continuous monitoring and guidance.
There are two questions about that stability I have.
One, things like catastrophic forgetting and falling into incoherence.
Two, less likely but far more worrying, falling into unwanted attractor states. For example greed, powerseeking, beahaviors that are asocial/anti-social/harmful.
It will eventually be super useful, and so disruptive that it will make today's LLMs look like nothing particularly special IMHO.
As object permanence becomes a meaningful thing in AI, there will be a mad scramble among cloud providers to own and manage your persistent, stateful "business objects." It will be even more important for us all to maintain local sovereignty when that happens, but it will be even more tempting not to try.
Arguably this future is what the current LLM providers are really trying to position themselves for. Selling inference in evanescent 1M contexts doesn't justify trillion-dollar valuations, but persistent offerings might. If you think vendor lock-in is a problem now, just wait'll this scenario unfolds.
Serving requests where every user has their own set of self-updating weights will absolutely murder whatever minimal margin the AI providers have today
Who says the weights have to be duplicated in their entirety? Even that will likely be worth it.
Imagine an OpenAI owning the ERP and CRM databases and workflows of a big chunk of the Fortune 500. Their typical customer's employee headcount might be 10% of what it once was, and OpenAI might capture 25% of the resulting savings. The contracts are signed in the same hemoglobin-based ink that Larry Ellison uses.
I see a new version of the web, web 4.0, being exactly this. Web 4.0 is the source of the live data. Web 4.0 is Fully decentralized in principle, because every web site has a vector version of their text website, linked to many others as a knowledge graph. The knowledge graph's links are like hyperlinks.
The vector databases might be free or paid or have ads. There is no difference between web 4.0 and an ai model that has all of it as its live data.
It is up to every client to turn that data into ai responses so in theory anyone would instantly have frontier scale ai which sounds as crazy as everyone having a computer, and the internet and the www in the 1980s.
It also, by coincidence, fully fulfils the goals of web 3.0 aka the semantic web.
Moderation of content and ranking content like a search engine would be left to the client. Or an instance like a fediverse mastodon instance or atproto with bluesky. There will be a few large models like web browsers or operating systems or social networks or fediverse instances today that will compile web 4.0 into a result in an ai agent like say Hermes and perform content moderation and ranking. You could in theory use a decensored model to perform crime like you can with the web today with Tor like selling drugs or piracy, with search engines that are not Google or defederated mastodon instances
Humans can't read raw html easily and can't read vector databases easily either. It would be read by an ai model connected to an ai agent, like a rendering engine connected to a graphical user interface
It might need a new protocol from the IETF and a standards body like the w3c? And a web server could automatically vectorize content via embeddings and link them in a decentralized knowledge graph for either graph rag after undoing vectorization or like this paper suggests being connected directly to the vectors themselves. The embeddings would be decentralized too or the website could offer them and there could be a way to translate between them
Web crawling would die and so would search engines because they would move to the user with their own ai compute and ai agent and data would already be machine readable and queryable. The data is always online and queryable programmatically so there would be no need for offline data silos. Unlike mcp servers which already do this, the data is not text based and there is no search engine to link pages to queries because the knowledge graph would be exactly that. The knowledge graph would be the new version of DNS for translation of a query into data like ip addresses into web domains because knowledge graphs are what search engines already are in a way so I don't see how a search engine would be useful except for a human to query it directly as done currently
now imagine that future frontier LLMs weights may be hard-wired in a chip (for performance & power efficiency), and any adaptations/tuning for them will be a blob of additional weights supplied by frontier labs (that will have to be in RAM)...
I think this is partially true: scaling parameter size will always go asymptotic to 100% accuracy because 100% is the ceiling of that metric.
However 95% is still half the error rate of 90%, and 97.5% is half the error rate of that.
And when test time compute like reasoning and looping harnesses stack many inference acts with many tokens each, those seemingly small accuracy gains stack tremendously.
Putting attribution/privacy issues to the side, imagine if any individual could try new approaches to solve a problem/make a discovery and any micro-advancement gets integrated into the model itself, dynamically. This could transform progress from the slow "write a paper, get peer reviewed and published, use published data to inform future work" to a system with a centralized repository of concepts, attempts and results, including failed approaches already tried. How much work do humans waste replicating failed approaches?
Someone completely random halfway around the world could trigger a prompt that solves a blocker that prevents my solution from working. Who cares about AGI or "can models invent anything" when we could have a system that automatically synthesizes individual human thought into a rich network of aggregate human memory.
That's the target OpenAI/Anthropic should be evangelizing, not an AI Daddy Overlord or agentic script kiddie hellscape.
It’s not the destination, it’s the journey to get there. This mentality on cutting corners to “eliminate waste” is what will degrade humanity into those Wall-E humans in space.
For example, if we just say, “Oh, someone did this already, why bother?” Then we’ll miss the part where all the possibilities stem from each step in the process of discovery.
Do what you love to do. Solve problems people have already completed. That’s where perspective comes from. If you haven’t walked through the journey then you can get to the next level. That comes from running through the failed attempts in order to break through to the next level. Don’t worry about what people have and haven’t done. If you’re doing something enjoyable, then that’s all that matters.
You'd want an enormous vector database at minimum. Text is just completely wrong for models at this scale, you must work in the latent space directly.
Every existing text web server can voluntarily offer a vector version of their website and charge for it or inject advertisements onto their content so the ai labs don't have to do it all themselves as model pretraining off a dataset. The vector version can have many links to other websites in the knowledge graph. This would be decentralized so not a monopoly and everyone not just ai labs would contribute to ai development because the dataset would be open because it would come from the internet itself as it already is (except for synthetic or user data).
They only ever placed them there because they saw it as unattainable. Rest assured that those posts will never stop moving.
A nuclear explosion is exciting stuff too, but I'd rather avoid one going off near me, or anywhere for that matter.
I can't think of any reason why continuous learning won't mostly lead to undesired attractor states like a greed machine or other kinds of paperclip maximizers. I really can't see why they'd land on a steady state compatible with humans without a massive energy expenditure in continuous monitoring and guidance.
They are unpredictable enough without learning, this is cool but I wonder how useful it will be in the long run
One, things like catastrophic forgetting and falling into incoherence.
Two, less likely but far more worrying, falling into unwanted attractor states. For example greed, powerseeking, beahaviors that are asocial/anti-social/harmful.
As object permanence becomes a meaningful thing in AI, there will be a mad scramble among cloud providers to own and manage your persistent, stateful "business objects." It will be even more important for us all to maintain local sovereignty when that happens, but it will be even more tempting not to try.
Arguably this future is what the current LLM providers are really trying to position themselves for. Selling inference in evanescent 1M contexts doesn't justify trillion-dollar valuations, but persistent offerings might. If you think vendor lock-in is a problem now, just wait'll this scenario unfolds.
Imagine an OpenAI owning the ERP and CRM databases and workflows of a big chunk of the Fortune 500. Their typical customer's employee headcount might be 10% of what it once was, and OpenAI might capture 25% of the resulting savings. The contracts are signed in the same hemoglobin-based ink that Larry Ellison uses.
The vector databases might be free or paid or have ads. There is no difference between web 4.0 and an ai model that has all of it as its live data.
It is up to every client to turn that data into ai responses so in theory anyone would instantly have frontier scale ai which sounds as crazy as everyone having a computer, and the internet and the www in the 1980s.
It also, by coincidence, fully fulfils the goals of web 3.0 aka the semantic web.
Moderation of content and ranking content like a search engine would be left to the client. Or an instance like a fediverse mastodon instance or atproto with bluesky. There will be a few large models like web browsers or operating systems or social networks or fediverse instances today that will compile web 4.0 into a result in an ai agent like say Hermes and perform content moderation and ranking. You could in theory use a decensored model to perform crime like you can with the web today with Tor like selling drugs or piracy, with search engines that are not Google or defederated mastodon instances
Humans can't read raw html easily and can't read vector databases easily either. It would be read by an ai model connected to an ai agent, like a rendering engine connected to a graphical user interface
It might need a new protocol from the IETF and a standards body like the w3c? And a web server could automatically vectorize content via embeddings and link them in a decentralized knowledge graph for either graph rag after undoing vectorization or like this paper suggests being connected directly to the vectors themselves. The embeddings would be decentralized too or the website could offer them and there could be a way to translate between them
Web crawling would die and so would search engines because they would move to the user with their own ai compute and ai agent and data would already be machine readable and queryable. The data is always online and queryable programmatically so there would be no need for offline data silos. Unlike mcp servers which already do this, the data is not text based and there is no search engine to link pages to queries because the knowledge graph would be exactly that. The knowledge graph would be the new version of DNS for translation of a query into data like ip addresses into web domains because knowledge graphs are what search engines already are in a way so I don't see how a search engine would be useful except for a human to query it directly as done currently
Which is a choice, not a "law":
https://arxiv.org/abs/2510.13786
https://www.alphaxiv.org/abs/2512.20264
https://arxiv.org/abs/2607.05155
However 95% is still half the error rate of 90%, and 97.5% is half the error rate of that.
And when test time compute like reasoning and looping harnesses stack many inference acts with many tokens each, those seemingly small accuracy gains stack tremendously.