So here's a big research bet that all the labs are making.

They think that if we train AIs to accomplish millions of verifiable

tasks across thousands of diverse RL environments, then we will have

basically built AGI, because this kind of training will have created a kind of

problem-solving agent: the kind of thing that can make progress on open-ended

tasks for weeks on end in the face of errors and mistakes and ambiguity.

And the people who are optimistic about this vision will say that all

these things we talk about as the fundamental deficits in the current

training paradigm — for example, the data inefficiency of these models,

or the fact that they lack continual learning — can just be steamrolled if we

scale training more, in the same way that all the fundamental research problems

in natural language processing collapsed when we threw enough compute into LLMs.

So in the previous essay, I talked about how these models are one

one-millionth as sample-efficient as humans, and the people who are in favor

of the current training paradigm will say, "Look, that might be true, but

this is only true during training."

Training is this one-time cost that is amortized across billions of

sessions that a model will experience.

What really matters is how smart and general and sample-efficient

the model is during a session, and this has clearly been improving as

we've been doing more RL training.

AI agents are able to solve more and more ambitious problems over

longer and longer time spans.

Anybody who has used these models for coding knows that.

Similarly, people would say, look, continual learning — this capability

I keep harping about, where the model's weights get updated based on

what it's learning from deployment — may simply not be necessary.

Because if in-context learning gets so good across longer and longer

time horizons, then you don't need to distill everything the model is learning

on the job back into the weights.

People often say that their employees are not net productive

until six months or more on the job.

So clearly, online learning is necessary for competence.

But what if you could just fit those six months into the context window?

There have been tons of architectural innovations that dramatically increase

the amount of information, or the amount of context, that a transformer can store.

And why not think that, with a couple more years of progress,

we might have what feels like infinitely large context windows?

Okay, so before we discuss this research bet a bit further, I want to step

back and ask a completely tangential question, which I find actually

very interesting and confusing about the nature of current AI progress.

Why has progress on computer use been so much slower than other domains?

Computer use is so clearly verifiable.

You could ask a question like: did the desired Etsy item I ordered get delivered?

Is the venue for an event I'm trying to organize booked?

Have my taxes been submitted?

So isn't it weird that computer use has been making so much slower

progress than coding and math and these other verifiable domains?

I'm sure there are many reasons for this, and one of them, of course,

is the fact that the models are exposed to far less high-quality

multimodal data during pretraining.

But one reason that I think is actually quite underrated by people, and which I

think reveals the canyon walls against which this river of AI progress will

only slowly chip away, is that it is not enough for a domain to be verifiable.

It also has to be very grindable, in the sense that you have to be able

to run lots of parallel rollouts against a deterministic and replayable

simulator, and you have to run those rollouts from the same starting point.

If you're trying to make a model better at coding, you can define some

container that has a software repo with some missing feature that you

have tasked the AIs with creating.

And then you have a thousand parallel agents go at the problem, each of which

has an identical copy of the container.

But this doesn't work with computer use, at least not trivially.

You can't just have a thousand agents go try the same checkout flow

on Amazon to get better at using websites, because Andy Jassy will

find your bots and shut your ass down.

You can solve this by making clones of Slack and Gmail and all the other

common applications and websites.

But at least currently, this is a very labor-intensive and unscalable

way to build environments.

Of course, once AIs get good enough at coding themselves to build these

clones with extremely high fidelity, then I'm sure computer use will make

quicker progress than it is right now.

And you're also killing two birds with one stone with this kind of procedure,

because getting AIs to rebuild whole applications from scratch is also

a great RL objective for coding.

So while computer use itself may soon be solved, its current lethargy

is telling us the following: that unless you can build a very replayable

training target for a domain, the models will struggle to make much progress.

And the reason this is true, of course, is that the models are incredibly

sample-inefficient during training.

This is the point I was making in my last video essay.

So for computer use, we might be able to make up for the sample-efficiency

deficit by building these farmable deterministic simulators.

But for so many other different kinds of skills that we need AIs

to have, we simply can't do this.

How do we train an AI to get really good at building a business from scratch?

How about winning court cases, or having a profitable day of trading in the markets,

or helping a candidate win an election?

The rollout here requires interacting with the real world, and you can't

recreate it from just within a datacenter.

The outer-loop verification here may take months or even years of

real-world actions to elicit, and you can't re-observe it by perturbing the

model's actions slightly in thousands of parallel rollouts to isolate exactly

what the model did that actually worked.

Now, dealing with such reset-free, non-stationary environments

is a known open problem in RL.

I'm not pointing out anything new.

But I really do want to emphasize that because of the idiosyncratic

and sparse nature of data in most domains in the world, you need sample

efficiency in order to get proficient.

If AIs are to develop all the skills that humans have, and even skills

that humans don't have, then they need to be able to learn from information

revealed in unstructured, unverifiable, and ambiguous ways from scarce

amounts of real-world interaction.

Because in many domains, the relevant training information simply

doesn't exist in any other way.

What is the RL environment to make an AI that is as good at politics as

Lyndon Johnson, or as good at building a space-launch business as Elon Musk?

The labs are betting that RLVR will generalize.

That is, that if you train on enough containerized, reproducible

environments, you will develop a very general agent that can make and

execute plans and learn rapidly from new information, and even pick up new

skills, all within a single session.

If you drop this endlessly RLVR'd AI into Texas politics in 1948, it

could give you better advice than LBJ about winning the Senate seat.

And if you gave it a hundred million dollars in 2002 and let it cook,

it would build SpaceX for you.

Now, whether RLVR can generalize this well is an empirical question.

If the labs went from spending billions of dollars on RL environments to a trillion

dollars, would you get the kind of thing that is a fully human-like general

intelligence within the context window?

Dario gave a telling quote during our podcast together, which I think

hints that RLVR generalization is not infinitely strong.

When he was explaining why model performance tends to degrade at long

context, he said: "There's two things.

There's the context length you train at, and there's a context

length that you serve at.

If you train at a small context length and then try to serve at a long context

length, maybe you get these degradations."

Now, maybe I'm reading too much into this, but it seems like he's

saying that short-horizon RL training doesn't necessarily generalize

to long-horizon RL performance.

And if you can't generalize from short horizon to long horizon, then how are

agents supposed to generalize from getting trained at a bunch of white-collar tasks

to, say, having the ability to be dropped in the real world and build a business

from scratch as well as Sam Walton?

And even if, after enough in-context experience, the AIs could become

like Henry Ford or Albert Einstein, all that would be ephemeral and

wasted if you couldn't get those learnings back into the weights.

Around 30 to 50 percent of a lab's compute goes to inference, and that compute is

currently not playing any productive role in helping improve the model.

This seems like a huge waste.

And it's even worse than it sounds, because it is only in deployment

that the most valuable bits of information which your model could

learn from are actually revealed.

What's actually happening in the organizations where I'm being used?

What are they using me for?

And what kinds of mistakes do I tend to make in the real world?

We've got some genius grad student who's never been allowed to take a real

internship, and we keep giving it more and more classroom case studies in the

form of RL training on environments.

It's so bizarre that we have AIs that are broadly deployed through the

economy already, and are participating in so many different kinds of tasks,

and are privy to so much domain- and organization-specific tacit knowledge,

and they're not able to make use of it.

But this kind of continual learning requires going back to the weights.

AIs can't just keep building up a bigger and bigger KV cache as they

learn from more and more users.

That's just not scalable, and that's also not how humans do it.

There's no clean separation in our brain between parameters and activations,

and it's not like some part of your skull keeps expanding as you learn

more things throughout your lifetime.

When we learn stuff, there's clearly some kind of compression, and this

aids our generalization and grokking.

There are, in fact, some humans who have this autistic-savant-type ability

to recall random tables of numbers or nonsense syllables years later

— basically the kind of fidelity of information that models have in context.

And such sheer volume cripples these humans' ability to understand

abstractions and metaphors.

Human continual learning is less about having all your observations at the tip

of your tongue and more about chiseling the right intuitions and big-picture

knowledge back into the weights.

But the moment you move into the weights, you have to give up on

in-context learning's sample efficiency.

Because gradient updates are super sample-inefficient, all of the

successfully shipped online-learning models have had to learn the exact

same thing across millions of users.

For example, the Cursor Tab model online-learns by predicting the

same exact objective for over 400 million requests a day.

The objective here is which edits actually got accepted by the user.

At least so far, we haven't seen models online-learn different kinds of things for

different users, because while a single session may generate more than enough

data for a human to learn from, it's not enough to train a more capable AI.

Current online learning can work for a very limited number of use cases.

But the whole point of continual learning is that the world is very

complicated, and each job and company and problem is different, and you need

your intelligence to be able to learn the specific information related to a

particular deployment, which simply can't be stuffed into some shared training run.

These are all the things we're talking about when we talk about on-the-job

learning: things like how everything in your organization works and fits

together, how to cooperate with all the infrastructure and the other people

around you to make progress on some larger project, what the common failure modes

are, and many other things like this.

As the podcast has grown, I've had to deal with more and more operational overhead.

Take paying bills.

In the past, contractors would just email me their invoices.

Every few weeks, I'd dig through my inbox, create a folder with all the

bills, and manually pay each one.

At this point, though, I just give everybody an email address

that goes straight to Mercury, which is my banking platform.

Whenever anybody sends an invoice to that address, Mercury automatically

downloads it, scans it, and extracts all the relevant information — things

like the contractor name, address, payment amount, invoice number,

and due date — and then uses all of this to create a draft payment.

Mercury then stores a list of these drafts for me to review.

I just go through the list and double-check that

they've been built correctly.

I don't have to track anything or enter any information myself.

Mercury does all the fundamental things for your business extremely well,

and it puts them all in one place.

If you want to learn more, go to mercury.com.

Mercury is a fintech company, not an FDIC-insured bank.

Banking services provided through Choice Financial Group

and Column N.A., Members FDIC.

In this way, sample efficiency and continual learning are actually

deeply connected problems.

Relatively little data is available to the model on the job.

Now, to learn from this data requires sample efficiency, and models can do that

in context, but using the fast weights that are built on the fly by attention,

which allow for this sample efficiency, scales very poorly in terms of memory.

So we need architectural innovations that allow for some kind of

intermediate representation.

I talked before about how we already have many different working ideas

for this kind of thing, from sparse attention to KV cache compaction.

And every week, somebody releases a new paper suggesting some kind of

other architectural optimization.

It doesn't seem to me that architecture is fundamentally what

is bottlenecking continual learning.

So perhaps the bottleneck is the loss function.

How do we update the weights, AKA how do we improve the model itself,

based on information that was learned from one particular session?

Even here, naively, it seems like there are many ideas that ought to work.

A lot of people are talking about this technique called on-policy

self-distillation recently.

If you want to learn more about it, I recorded a little impromptu

blackboard lecture on my iPhone with Sasha Rush a couple weeks ago, and

it's in the link in the description.

But to summarize the explanation, the idea is that we encourage the base model

to make the same predictions when trying to solve some real-world problem as the

model with all the context accumulated after a long session would have made.

The whole point of this procedure is to distill what the model learned in a

session back into the weights themselves.

This is better than RLVR for two reasons.

One, OPSD doesn't require us to have some outer-loop verifiable reward.

We just need a model that can learn the right things within the context window.

And as long as we have that, we can train the base model to match our veteran

teacher model, which has built up all this experience during the session.

And two, OPSD provides a much denser supervision signal than naive RL.

Instead of projecting a single reward through the whole

trajectory, you can train on the per-token probability discrepancy

between the teacher and student.

For continual learning, OPSD is also superior to supervised fine-tuning.

The most naive version of SFT for this application that you can

imagine is just to train the base model to predict all the tokens that

are observed during the session.

But this makes no sense if you think about it as a learning target.

The way you get better at your job is not by recalling the transcript

of every single thing that happened every day with perfect fidelity.

Rather, it's by consolidating the handful of insights and pieces of

knowledge that are actually relevant to you getting better at your job.

RL training doesn't suffer from this failure mode.

RL is great at concentrating the update to only what is relevant

to getting the outcome right.

That's why

the updates from RL are incredibly sparse.

And this is a very important property for continual learning, because as

you're learning on the job, you don't want to overwrite and forget all the

other things that the base model knows.

I wrote a post a few months earlier arguing that RL learns

much less information per sample than supervised learning.

But this may be a good thing rather than a bad thing.

You only change the model as much as is absolutely necessary to

achieve the outcome, and no more.

OPSD preserves this property of RL, where instead of slingshotting towards

the teacher distribution as supervised learning would have you do, you only

extract the knowledge that is necessary to achieve the same results as the

teacher on actual real-world tasks.

OPSD is one way to attack the sample-efficiency problem.

You take this scarce real-world experience, and you squeeze all the

signal into a tiny, well-targeted update.

But there's also another much more speculative idea.

Let's call it dreaming.

If the AI can build a good simulation of reality against which to rehearse

new skills, or try alternative strategies and reinforce what actually

works, then AIs could experience orders of magnitude more simulated

samples in the same wall-clock time.

Let's go back into history a bit.

A couple years after DeepMind released AlphaZero, a group of researchers

trained a model called EfficientZero, and the whole point of this model

is to be very efficient with data.

So if this model and a human both got two hours to play against a simulator

of an Atari game that they hadn't seen before, this model would actually

probably beat the novice human.

Does this mean that the model was more sample-efficient than the humans?

Well, that was the goal of the training, but it depends on how

you measure sample efficiency.

Because for each step in the real game, EfficientZero is playing dozens

of simulated games in its head.

In a similar way, future LLMs might be able to consume far less

real-world data while practicing endlessly against environments

that they build for themselves.

The big difference, of course, is that it will be much harder to build

a simulation of the whole world than it is to emulate the game of Go.

That's why I said this is a much more speculative idea.

If it works, it would become a fourth axis of scaling alongside pretraining,

RL, and inference-time compute.

You could call it test-time training or dreaming.

The model spends compute writing up RL environments and then training

against them, and it's rehearsing all the skills that will actually be used

in production for a specific user.

So instead of hitting /compact in Codex or Cursor or Claude, which kindles a small

amount of compute to write up a summary, and which gives you the simulacrum of

continual learning, you hit /dream.

And this incinerates huge amounts of compute to build and train against a

video-game version of what the model is witnessing in the real world.

So what might continual learning look like by 2027 or 2028?

And how do we get there?

Here's one scenario.

All of this RLVR training is producing an agent that can get its bearings when

it's thrown at an unfamiliar problem, and it can try different strategies, and

it can iterate when it hits a roadblock.

This is the crucial thing that RLVR has given you: an AI that

is at least competent enough to start getting some real-world

experience, if it could learn from it.

And once you have that, you send it out into the world to do real

work, even on projects that are off the training distribution.

Now let's say at this point, the effective context lengths have expanded

such that AIs can jam and co-work with you for a full week of wall-clock time.

At the end of a week, you give it a thumbs up or a thumbs

down, you give it a work review.

And if you give it a thumbs up, the base model distills everything that

the AI learned during the session, and it may use OPSD, it may use dreaming,

it may use some other technique that we aren't even aware of, or it'll use

a combination of all of the above.

And AI can get better at domains that are adjacent to what it was explicitly

trained for beforehand with RLVR.

And in the next round it gets better at the thing adjacent to what it

was previously online learned.

In this way, the gamut of AI skills and knowledge and capabilities can

expand far beyond the verifiable domains that the model was originally

trained against before it was deployed.

Just as pretraining created a base intelligence that was smart enough to

become a competent agent with enough RLVR on top, so RLVR has created an agent

that is competent enough to actually be broadly deployed in the world, and

from this broad deployment to learn on the job once the training recipe for

continual learning actually arrives.

By this point, the main way that AIs get better is not from the

training they have received before they are released to the public.

Rather, it's from all this experience that they'll be accumulating from being broadly

deployed in the economy and engaging in so many different kinds of tasks.

Every time that you interact with an AI, it'll be smarter, not only because

it's been learning from your previous sessions, but also because it's been

learning from all its interactions with all the other users in the world.

And that's very scary and exciting and different from the

way that AI improves right now.

This was a narration of a blog post that I also released on

my website at dwarkesh.com.

Go there if you want to read all the footnotes, or if you want to

sign up so you can find out when I release the next blog post.

Otherwise, I'll see you on the next episode.