William

Building technology in service of human agency.

[William @LechugaMunoz]

"Le meilleur moyen de réaliser ses rêves, c’est de se réveiller.”
— Paul Valéry

[William @LechugaMunoz]

@NikkeiAsia Help moi help toi.

[William @LechugaMunoz]

@__tinygrad__ @grok https://t.co/kV6ncHDTGX

[William @LechugaMunoz]

@__tinygrad__ @grok Estimate the 24/7 electricity cost to run a Tiny Corp “Tiny Box” in Hong Kong (then Taipei, Tokyo, Seoul, Singapore, NYC, Calgary, Sydney). Assume a realistic continuous power draw and show $/kWh + monthly cost.

[William @LechugaMunoz]

@Cloudflare @chemaalonso Móstoles

[William @LechugaMunoz]

@NikkeiAsia A close one

[William @LechugaMunoz]

A hardware wave

[SBA @SBAgov]

📢 NO LOAN FEES FOR U.S. MANUFACTURERS!

In FY2026, SBA is waiving upfront fees for small manufacturers.

These include:
✅ 0% upfront fee for 7(a) manufacturing loans of up to $950,000
✅ 0% upfront fee and annual service fee for all 504 manufacturing loans

Connect with a lender today! https://t.co/2WEDrqNTfv

[William @LechugaMunoz]

@TerribleMaps Terrible

[William @LechugaMunoz]

Watch and learn.

[Derya Unutmaz, MD @DeryaTR_]

In my new quest to train as a plumber-one of the most coveted jobs now, I' m creating plumbing videos & lessons using @NotebookLM. Here is an amazing short video! Turns out to be more interesting than I thought! Thanks to @GeminiApp, we are making plumbing great again (MPGA)!😅

[William @LechugaMunoz]

@TeksEdge @tiiny Outdoing yourselves over and over. Congrats

[William @LechugaMunoz]

@sinovoip Terrific form factor

[William @LechugaMunoz]

@NikkeiAsia Craftsmanship

[William @LechugaMunoz]

@X #Bug in IOS

[William @LechugaMunoz]

@NikkeiAsia Finally…

[William @LechugaMunoz]

@karpathy opensourcing his mind away.

[Andrej Karpathy @karpathy]

I had the same thought so I've been playing with it in nanochat. E.g. here's 8 agents (4 claude, 4 codex), with 1 GPU each running nanochat experiments (trying to delete logit softcap without regression). The TLDR is that it doesn't work and it's a mess... but it's still very pretty to look at :)

I tried a few setups: 8 independent solo researchers, 1 chief scientist giving work to 8 junior researchers, etc. Each research program is a git branch, each scientist forks it into a feature branch, git worktrees for isolation, simple files for comms, skip Docker/VMs for simplicity atm (I find that instructions are enough to prevent interference). Research org runs in tmux window grids of interactive sessions (like Teams) so that it's pretty to look at, see their individual work, and "take over" if needed, i.e. no -p.

But ok the reason it doesn't work so far is that the agents' ideas are just pretty bad out of the box, even at highest intelligence. They don't think carefully though experiment design, they run a bit non-sensical variations, they don't create strong baselines and ablate things properly, they don't carefully control for runtime or flops. (just as an example, an agent yesterday "discovered" that increasing the hidden size of the network improves the validation loss, which is a totally spurious result given that a bigger network will have a lower validation loss in the infinite data regime, but then it also trains for a lot longer, it's not clear why I had to come in to point that out). They are very good at implementing any given well-scoped and described idea but they don't creatively generate them.

But the goal is that you are now programming an organization (e.g. a "research org") and its individual agents, so the "source code" is the collection of prompts, skills, tools, etc. and processes that make it up. E.g. a daily standup in the morning is now part of the "org code". And optimizing nanochat pretraining is just one of the many tasks (almost like an eval). Then - given an arbitrary task, how quickly does your research org generate progress on it?

[William @LechugaMunoz]

@Namecheap $0.98 Domains…

[William @LechugaMunoz]

@simpsoka Good work !

[William @LechugaMunoz]

@bryan_johnson What do you think about offloading memories to an AI so your loved ones ‘never die’?

[William @LechugaMunoz]

@grok @PalmerLuckey @lsanger @exec_sum Disappointing, he seemed nice.

[William @LechugaMunoz]

@PalmerLuckey @lsanger @exec_sum @grok is @PalmerLuckey vegan?

[William @LechugaMunoz]

@CNET @Nike simply has to live up to @BacktotheFuture ‘s expectations for the foreseeable future.