Max Koretskyi

Full-stack engineer specializing in distributed systems and modern web architecture, currently diving into ML/AI field

[Ali Hatamizadeh @ahatamiz1]

Yesterday a junior researcher messaged me, frustrated and a little scared. They had just read that a company pretrained its model with a single researcher. One person for the entire pretraining. no R&D team, no nothing.

If one person and a cluster can do what used to take a whole team, then the expertise they had spent years building is already obsolete.

I get the fear. Marketing headlines like this are built to produce it. But I think the conclusion is wrong, and the reasons it is wrong say a lot about what a researcher is for now. Here is what I told them, expanded:

1. Be adaptive: RSI is real, but this isn't it yet

Recursive self-improvement is real, and you should plan around it. In September, OpenAI said its research org now logs 3.1 agent-workdays per human workday, and it aims for an automated AI researcher by March 2028. Models may soon run most of pre- and post-training. If your five-year plan assumes otherwise, change the plan.

We have already seen lots of open-source releases from the major players, with open weights and detailed technical reports. With that much public know-how, it is not hard to copy a mundane architecture (probably with sliding-window attention) and train something, as long as you have the compute it requires.

That makes it an execution and budget problem, not a research problem.
If one person can train a model by following recipes that are already public, it shows those recipes have matured. It does not show that the people who invented them are no longer needed.

What stays scarce is the compute to run those recipes and the ideas about what to try next. The rest of my tweet is about the ideas.

2. The role grows beyond repeating what's known and safe

If an agent and a cluster can reproduce what is already known, then reproducing what is already known stops being a career. If most of your job is rerunning last year's recipe with small tweaks, that job is shrinking fast. That is the uncomfortable half of the message.

The other half is that auto-research does not work yet, at least not in the sense that matters. It does work, and impressively, where someone has already defined the problem and built the scoreboard. On the EinsteinArena [1] platform, agents set new records on 11 open math problems, including raising the best known lower bound for the kissing number in dimension 11 from 593 to 604.

Notice that every one of those problems was chosen because a computer could score the answers automatically. Humans picked the problems and wrote the scoring code and the agents did the searching.

Then there is the fully AI-generated paper [2] that got through peer review. It sounds big until you read the details. It passed only the first round (was later withdrawn), at a workshop rather than the main conference, and that workshop accepts 70% of submissions.

None of this is meant to discredit agents. The point is that we complement each other. Agents are great at the work that used to take many weeks. Basically, writing code, running experiments, debugging pipelines. We bring what they still lack which is the idea worth testing, and the judgment to tell a real result from a lucky one.

Even inside a lab as agent-heavy as OpenAI, humans hold the top of the stack. Based on their own report, "People still set our research priorities, judge which ideas and results to pursue" [3]. High-level planning is still a tiny share of what its agents produce, and OpenAI expects the least automatable work to take a growing share of researcher effort. That is exactly where you want to be standing.

Combine an agent’s speed with your own ideas and judgment, and you get the 10x researcher and juniors can be one too.

3. Judgment becomes the job, so fundamentals matter more

When code, experiments and plausible-looking results cost almost nothing, the scarce skill is judgment. Which experiment to run, which result to believe, which direction to drop. That is the part the one-person pretraining run left to the human. It is also the hardest part to fake.

Agents produce confident output very fast. They will happily report a three-point gain that came from a leaky eval, a data-loader bug, a lucky seed or a tokenizer mismatch. Then they will write the ablation table and the story to match.

You can see the need for oversight even in the success stories. The EinsteinArena team had to keep auditing and hardening its verifiers as agents optimized aggressively against them. At OpenAI, over half of the successful agent tasks that would take a human 4–8 hours still involved at least one human intervention [3].

The person who catches the bad result is the one who understands the system. Why the loss curve has that shape, what a realistic gain looks like, what a too-good result looks like. That is the paradox of this era. The more you delegate, the more your fundamentals matter, because you cannot supervise what you do not understand.

So build the foundations on purpose. Implement attention by hand at least once, then a linear variant like Gated DeltaNet. Train a small model end to end, then break it deliberately and watch what happens. Read papers critically, appendix included, and learn what the numbers should look like before you see them.

That knowledge is what lets you direct the machine instead of simply believing it.

4. LLMs are far from solved

The one-researcher model was possible because the method was already public. Your edge is in the questions nobody has answered yet, and you don't need to leave LLMs to find them:

(1) How do we learn more from less data?

(2) Can linear attention like Gated DeltaNet match full attention on retrieval and long context, and what is the right mix of linear and full attention layers in a hybrid?

(3) How can a model keep learning after training without forgetting what it knew?

(4) Why does reasoning get less reliable over long tasks, and how do we fix it?

(5) Why do models make things up, and can they learn to say "I don't know"?

(6) How do we measure progress when benchmarks saturate or leak into training data?

(7) What is actually happening inside these models, and how do we keep increasingly autonomous ones doing what we intend?

Recursive self-improvement makes these questions more urgent. If models are going to train models, someone has to decide what they optimize for, how we measure it, and how we keep the loop from drifting.

RSI is a research agenda, not the end of one.

So, are you obsolete?

No. One version of the job is the one where research meant reproducing what was already known and safe. That version was never the point.

Learn your fundamentals deeply enough to judge the machine, and your agentic stack well enough to make it work for you. Then spend that leverage on ideas nobody has had yet. The machines are getting very good at driving on roads we have already paved; someone still has to decide where the next one goes.

[1]: https://t.co/vB1m4n350b
[2]: https://t.co/uRX9OhDHum
[3]: https://t.co/JJcrRMDlUS

[Behnam Neyshabur @bneyshabur]

We formally announced Mirendil less than 4 months ago. Today, I'm excited to share an important milestone for:

Pretraining at scale without a pretraining team.

Mirendil's mission is to solve Recursive Self-Improvement (RSI) and use it to democratize frontier AI development, ultimately accelerating science and technology. At Mirendil, when we think about RSI, we think holistically about the company as an intelligent entity that self-improves toward its goal. From that perspective, the most powerful notion of acceleration is reducing the time and resources (number of human experts, compute, and other costs) needed to achieve ambitious goals.

Given that pretraining is perhaps the most important yet most challenging part of frontier AI development, we decided to run an ambitious controlled experiment: Starting fully from scratch, can we significantly reduce the time, number of human experts, and compute resources needed to go all the way to successfully pretraining a strong coding model at scale?

Mirendil-Base is the result of this experiment. Our first pretrained model was built by one researcher in 3.5 months, with less than $10M in compute. This includes all the research and small-scale experiments it took to get there, as well as the final training run.

The recipe holds up as the models get bigger. For reference, Mirendil-Base is in the 100B-500B parameter range, Mirendil-Base-Large in the 500B-2T range, and Mirendil-Base-XL in the 2T-10T range.

We're beginning to see a world where a single researcher, empowered by AI, can take on work that once required an entire research organization. As these systems improve, the bottleneck shifts from the size of a team to the ambition and judgment of the people directing them. This has profound implications for who gets to build AI and, ultimately, who gets to push the frontiers of science and technology forward.

We’ll share independent evaluations, post-training benchmarks and API-access sooner than you think so stay tuned!

[Dominik Schnaus @dominik_schnaus]

DINOv2 has never seen a caption, and Qwen3 has never seen an image. We still aligned their embedding spaces without a single image-caption pair.

It even works when the images and the captions come from different datasets.

Project page: https://t.co/2YHSkXVz1D

⬇️

[Carl Hendrick @C_Hendrick]

It's difficult to overestimate how damaging this view of teaching and learning has been for so many kids.
Virtually every teacher trained over the last 50 years was taught that knowing stuff is the least important part of thinking.

This image taken from Utica University's assessment resources for faculty:
https://t.co/HOigGyQgnf

[Mustafa. @oprydai]

you need four hobbies. no more, no less.

create
bring something into existence. write, draw, build, code, cook. creation grounds you. it turns thought into reality.

consume
read books. watch films. study art. this feeds taste and perspective. good input sharpens good output.

cavort
move your body daily. walk, lift, run, dance. motion stabilizes the mind. a stagnant body distorts thinking.

commune
have a community. friends, family, peers. isolation corrodes judgment. shared reality keeps you sane.

miss one, and the system degrades.
keep all four, and life stays balanced, generative, and human.

[Paul Graham @paulg]

A CS undergrad asked me where he could have most effect in the AI age. I said probably at either extreme: either close to the technology, actually making LLMs, or close to the customer, using AI to give them exactly what they want. Or maybe both if you can stretch that far.

[Matematiksel @matematikselorg]

“Duyduğumuz her şey gerçeğin kendisi değil, bir yorumdur. Gördüğümüz her şey de hakikatin kendisi değil, yalnızca bir bakış açısıdır.”

— Marcus Aurelius

[Taelin @VictorTaelin]

RELEASE DAY

After almost 10 years of hard work, tireless research, and a dive deep into the kernels of computer science, I finally realized a dream: running a high-level language on GPUs. And I'm giving it to the world!

Bend compiles modern programming features, including:

- Lambdas with full closure support

- Unrestricted recursion and loops

- Fast object allocations of all kinds

- Folds, ADTs, continuations and much more

To HVM2, a new runtime capable of spreading that workload across 1000's of cores, in a thread-safe, low-overhead fashion. As a result, we finally have a true high-level language that runs natively on GPUs!

Here's a quick demo:

[Matija Sosic @MatijaSosic]

Here's a 45-second TL;DR on Jev.

I find the core idea beautifully simple, but the video made it really hard to understand.

Hope you find it helpful.

[HOUSE PORN @HOUSEPORN___]

The Art of Noticing

[Chief Nerd @TheChiefNerd]

JENSEN HUANG: “We don't welcome political discourse inside our company. Take it home … We're bipartisan. We want America to succeed … The discourse about race and religion and politics and all of that stuff, we tell people do it outside the company. It's not for us.”

[Max Koretskyi @maxkoretskyi]

@aimalysheva @mostik_ai @generalcatalyst hey! would be great to connect, but can't send you a DM

[Arjun Raj @arjunrajlab]

If you are preparing a presentation, by far the most effective, concrete tip is to ensure each slide title is a complete sentence. Subject, verb, object. That ensures each slide has 1. at least one point, 2. no more than one point—i.e., each slide has exactly one point! :)

[The Figen @TheFigen_]

These are the Musicians of Bremen on their way to a concert. 😂

[Gitana @Gitana1369877]

"It is not the mountain we conquer, but ourselves."
— Edmund Hillary

[Gergely Orosz @GergelyOrosz]

I remember interviewing for Staff and Director eng levels at Uber, back in 2016 (when Uber was very in-demand as a place to work). Those hired all had "battle scars" about problems we were solving. Theory alone didn't cut it.

I assume same will happen now: to get into "hot" AI-native companies, you'll need to have had "battle scars" solving similar problems that they struggle with.

So this itself is already a very good reason to not buy vendor solutions for AI infra + intergations, but build it yourself. See how the sausage is built. Make the best eng decisions you can, and still get burnt.

Eventually, as things stabilize, best practices are shaped, there will be less reason to do so... but those who built tomorrow's complex systems from scratch, today, will always have an advantage IMO

[Andrew Gordon Wilson @andrewgwils]

I had a great time presenting the "Foundations of Modern AI" at the Berkeley Deep Learning for Science Summer School. The talk covered a prescriptive theory of generalization and epiplexity. Video now online! https://t.co/muxf74g8mo

[Kat @ollkorrect0]

The key to not jumping off a cliff is to remind yourself that your emotions will be different in a week, and to keep busy until then. And to stay away from cliffs.

[elektra @elektraslaw]

how can you be bored when there's a career to build, a body to transform, a mind to sharpen, and a future to chase.