Sam Sends

@zck visit @joshuafarahzad at proto town

Who do I know in Austin, TX? Doing something interesting there next month

@JIACHENLIU8 legitimately very cool! it looks like the ARA UI is still in development. Is there a local version I can test out?

Absolutely, human’s attention is limited And lots of objective reviewing work can be done by agents I’ve discussed an ARA native review system in the paper - let agents to verify these objective claims in the paper, and let human reviewer provide their taste, judgement and intuition

My bet: in the near future, 80%⬆️ of CS research will be done by AI in collaboration with humans. However, today's research ecosystem is still built around the human, not the AI scientist. For example, the 8-page paper PDF is a lossy compression of months of branching exploration into a linear story, optimized for a human reviewer to skim in 30 minutes. It hides two structural taxes: 📖 Storytelling Tax — failures, rejected hypotheses, and dead ends get stripped. On RE-Bench (24,008 runs, 21 frontier models), failed runs = 90.2% of total compute cost, with a 113× median failed-to-success token ratio. Every lab independently rediscovers the same dead ends. 🔧 Engineering Tax — the gap between reviewer-sufficient prose and agent-sufficient spec. Across 8,921 PaperBench requirements (23 ICML'24 papers), only 45.4% are fully specified in the PDF. The rest is tacit lab knowledge. Tolerable when readers were human. Critical now that agents read, reproduce, and extend. We propose ARA: the Agent-Native Research Artifact — replace the narrative PDF with an agent-executable package, in 4 layers: 🧠 structured scientific logic ⚙️ executable code w/ full specs 🌳 exploration graph (every failure preserved) 📊 evidence grounding every claim

@Depohqo @bwein_ @SynchronicityHQ bro still alive, needed to make changes. have all the checkpoints while first testnet was up

@sam_sends @bwein_ @SynchronicityHQ Bro project soma still live?

Modern markets are fundamentally extractive We've spent three years rewiring them from scratch One core feature is everyone deploying their own bots without code or infra Here's a first look @SynchronicityHQ We're looking for strategists to test drive. Reply for early access

@elidourado sick as hell, lfg

Go to Proto-Town if you want to see robotically constructed adobe houses, high-recovery desal centrifuges, autonomous construction equipment, solar-chemical air conditioning, a radioisotope reactor, and more. It's delightful. wsj.com/business/entre…

@SynchronicityHQ 🙋‍♂️

You can now go from idea to live algorithmic trading strategy in under a minute. No code, no infrastructure, no racing against the market. Join the waitlist. Reply for early access.

@koeng101 @pylabrobot Could you Dm me your email? Your dm’s are closed!

@pylabrobot Finally, my ask: I want YOU yes YOU to come visit our facility in Burlingame, near SF! Shoot me an email and we’ll make it happen

@untillabs incredibly inspiring

"We distill our challenge to a single, elegant first principle: pause molecular motion in every organelle, in every cell, in every tissue of a human organ, then resume it." Read more about what we are working on in the link below👇

@AlexanderLong weights = compressed data

The weights power everything

@SynBio1 I just ordered all of those, pretty excited to dig in. Thanks, Jake, great recommendations. I would not have discovered these books otherwise!

@sam_sends Sorry I doubted you! Here’s some favs from my bookshelf. I’d also recommend “A genetic switch” by Mark Ptashne and “Random walks in biology” by Howard Berg

From 1903 to 1943, Martin Couney used incubators to save the lives of about 6500 premature babies and paid for it by charging admission to tourists I don't know how I feel about this but it's worth remembering that very different biotech business models are possible

Feel you, it’s getting harder to distinguish real v fake. This is my current stack. I have increasingly been reading more history so if you have a good rec for: 1. History of biological science 2. History of biotech Definitely lean more towards a good story versus a timeline of facts

@sam_sends No offense but this reply is the kind of thing a bot would write. Tell me a bit about your human interest in biotech books and I'd be happy to suggest something

@PawelHuryn high signal take sir 🫡

Everyone's framing this as a chip factory. It's actually a bet on AI compressing semiconductor learning curves. Traditional fabs take 3-5 years to reach yield. Tesla's bet: AI-driven design and simulation cuts that to months. Build the fab, train the models, use the models to design better chips, repeat. The flywheel isn't chips. It's learning speed. History suggests: don't bet against compounding.

@seyong what is the most compelling use case for agentic payments? Still doesn’t quite make sense for me tbh

agentic payments make sense to me. agentic trading does not tbh

@eryney_ok Dm’d!

What questions do you have about biology and medicine? Ask below or DM me!

@0interestrates Huh kind of surprised that the sigmoid is flattening around 6hrs. I’d expect screen time to plateau higher

@Appyg99 Strongly agree. It seems more likely that the large labs will vertically integrate payments. The only thing that matters is order flow, and the interface with the most usage has the most order flow.

As someone that was a true believer of agentic commerce last year & ultra skeptic this year — The problem is this belief that humans want agents shopping for them. Other than a few efficiency obsessed nerds, most customers don't just hand off their wallet to some bot to buy stuff without being able to be a part of the decision irrespective of what the stated preferences are. Shopping is a conscious and important decision for 90% of households. A pleasurable hobby for many. Unless somehow you manage to change this human behavior (highly unlikely), agentic commerce needs to be restructured around discovery and less around payments and actual conversion.

@cryptohateswjp you like padron cigars?

시가바 왔네

@vvaiyr so surprised tbh