Joon Yun

Physician. Hedge fund manager.

[Joon Yun @drjoonyun]

@paradeevic @grok who is the Harvard psychology professor in the video?

[Joon Yun @drjoonyun]

Now’s as fine a time as any, then, to take stock of those who might yearn for the same fair shake from us that we hope for from our putative robot overlords. After all, humans have turned a not insignificant portion of the biosphere into objects of husbandry.

[Joon Yun @drjoonyun]

AI Alignment—the process of aligning AI to human values—exhibits the recurring undertow of anthropocentricism. How about Generalized Alignment among sentient and living systems? Bring the Golden Rule to all?

[Joon Yun @drjoonyun]

For those used to being part of the hegemon, the prospect of slipping into the infraposition can be unnerving. Notice the angst over AI Hegemony and the mad scramble for AI Alignment.

[Joon Yun @drjoonyun]

Flying under the radar this week—tomorrow (June 29, 1776) marks the 250th anniversary of the founding of San Francisco

[Joon Yun @drjoonyun]

2013: The great Cowboy Ventures founder Aileen Lee coins the term “unicorn” to refer to the elusiveness of startups that reach the one billion dollar valuation mark.

2026: “kilocorn”

https://t.co/FEoOfFImyY

[Joon Yun @drjoonyun]

World went from rising temps are an existential threat to let's build existential threats while raising temps like never before

[Joon Yun @drjoonyun]

Full name on his driver's license: Eduard Karl Joseph Michael Marcus Antonius Koloman Volkhold Maria Habsburg-Lothringen

[Joon Yun @drjoonyun]

Epic reply by an outbred descendant of the House of Habsburg-Lorraine

[Joon Yun @drjoonyun]

https://t.co/G1SExU89Kk

[Joon Yun @drjoonyun]

"Punctuated Equilibria" of Agentic Evolution: Seen from an evolutionary lens, sharp inflection points of agentic evolution in @Karpathy’s graph may represent the agentic lineage crossing adaptive valleys in the landscape, with validation loss being the fitness function.

[Andrej Karpathy @karpathy]

Three days ago I left autoresearch tuning nanochat for ~2 days on depth=12 model. It found ~20 changes that improved the validation loss. I tested these changes yesterday and all of them were additive and transferred to larger (depth=24) models. Stacking up all of these changes, today I measured that the leaderboard's "Time to GPT-2" drops from 2.02 hours to 1.80 hours (~11% improvement), this will be the new leaderboard entry. So yes, these are real improvements and they make an actual difference. I am mildly surprised that my very first naive attempt already worked this well on top of what I thought was already a fairly manually well-tuned project.

This is a first for me because I am very used to doing the iterative optimization of neural network training manually. You come up with ideas, you implement them, you check if they work (better validation loss), you come up with new ideas based on that, you read some papers for inspiration, etc etc. This is the bread and butter of what I do daily for 2 decades. Seeing the agent do this entire workflow end-to-end and all by itself as it worked through approx. 700 changes autonomously is wild. It really looked at the sequence of results of experiments and used that to plan the next ones. It's not novel, ground-breaking "research" (yet), but all the adjustments are "real", I didn't find them manually previously, and they stack up and actually improved nanochat. Among the bigger things e.g.:

- It noticed an oversight that my parameterless QKnorm didn't have a scaler multiplier attached, so my attention was too diffuse. The agent found multipliers to sharpen it, pointing to future work.
- It found that the Value Embeddings really like regularization and I wasn't applying any (oops).
- It found that my banded attention was too conservative (i forgot to tune it).
- It found that AdamW betas were all messed up.
- It tuned the weight decay schedule.
- It tuned the network initialization.

This is on top of all the tuning I've already done over a good amount of time. The exact commit is here, from this "round 1" of autoresearch. I am going to kick off "round 2", and in parallel I am looking at how multiple agents can collaborate to unlock parallelism.
https://t.co/WAz8aIztKT

All LLM frontier labs will do this. It's the final boss battle. It's a lot more complex at scale of course - you don't just have a single train. py file to tune. But doing it is "just engineering" and it's going to work. You spin up a swarm of agents, you have them collaborate to tune smaller models, you promote the most promising ideas to increasingly larger scales, and humans (optionally) contribute on the edges.

And more generally, *any* metric you care about that is reasonably efficient to evaluate (or that has more efficient proxy metrics such as training a smaller network) can be autoresearched by an agent swarm. It's worth thinking about whether your problem falls into this bucket too.