Wicha (@flarrain) profile picture

I have designed quantum computing gates and built dynamic pricing systems at scale in the past. Leading the INCREDIBLE engineering team at Topsort

Palo Alto, CA
Joined June 2008
Photo by Wicha (@flarrain)Photo by Wicha (@flarrain)Photo by Wicha (@flarrain)Photo by Wicha (@flarrain)Photo by Wicha (@flarrain)
Wicha retweeted
Very happy to see this from Nick!
interpretability might be the most important scientific problem in the world right now, as I don't think we can solve alignment if we don't know what we're building. I think I have good news about it. I co-authored early work in mechanistic interpretability at OpenAI and came away pessimistic, because neural networks appeared to fold into arbitrary tangled knots. years later, I started openai's first large-scale automated interpretability project, finding about 2% of neurons legible with the primitive tools we had with modern agentic tools and a much closer look, the picture looks totally different. the vast majority in models up to trillions of parameters are legible, and their circuits beautiful. many famous examples of tangled neurons turned out to be misunderstood. The roadblocks I expected mostly aren't there. I now think ambitious interpretability will be solved by the time we need it most, and possibly quite soon, as automated scientists keep improving. it may be a bitter lesson for interpretability: the simple thing, at absurd scale, works already-clean.goodfire.com/
281
144
24
3,820
926,904
Wicha retweeted
The best predictor of success for tech companies, at every stage from during the YC batch to public company with billions in revenue, is the rate of shipping new stuff.
237
496
111
6,342
421,156
OpenClaw is the fastest-growing open source project, but there are no stories of running it safely in production at scale. As we started deploying agents internally at @brexHQ, we couldn’t stop thinking about this question. Agents work, but nobody wants to give them real credentials. Instead of waiting for a solution to emerge, we decided to try a novel approach: using LLMs to judge the network traffic of an AI agent. Today we’re announcing CrabTrap, an open-source proxy that intercepts every outbound request and blocks risky activity using LLMs, before it ever hits an external API. The results are promising; we believe it’s a meaningful step forward in the security of agent harnesses in production environments. Try it out today. (As a side note, it was really fun to work personally on a real systems problem again. And btw, if you want to work at a place where the CEO is building proxies at night, we’re hiring!)
102
146
44
1,911
776,531
Wicha retweeted
GBrain v0.10.0 is a big one My personal OpenClaw setup and brain can now be yours. I've perfected my RESOLVER.md, my SOUL.md and ACLs for multi-user brain access. Now there are 24 distinct fat skills with fat code, fully tested with e2e tests, evals and unit tests.
105
137
14
1,658
175,435
The "AI job loss" narratives are all fake. AI = massive ramp in productivity = massive ramp in demand = massive jobs boom. Watch.
1. Employer recovery from post-COVID hiring correction 2. Employer recovery from post-COVID interest rate spike 3. Elasticity = demand boom QED
644
778
248
7,994
1,283,293
6. The Influence of Kaiming He Xie deeply admires Kaiming He, with whom he worked closely at FAIR. He learned that Kaiming's superpower is his extreme focus and ability to build an impenetrable baseline ("scaffold") before aiming for breakthroughs [02:34:30]. Kaiming taught him that the real research "signals" come from the failures and surprises during the experimental process, not from armchair theorizing [02:08:12]. 7. "Impact" vs. "Understanding" Xie dislikes the word "impact," finding it too ego-driven and aggressive. Instead, he views research through the lens of philosopher Hannah Arendt: the goal of research is "understanding." Publishing a paper is about sharing a profound realization with the world to find resonance and intellectual kinship, rather than just changing the world by force [01:31:27]. 8. Intelligence is More Than Human Language Xie cautions against human arrogance in defining intelligence solely through language. Drawing on evolutionary biology, he argues that building an AI with the survival and physical reasoning skills of a squirrel is actually a much harder problem than building an AI that can write code or pass exams [06:13:05]. 9. The Power of "Research Taste" Xie equates research to filmmaking, where "taste" dictates everything from the problem you choose to solve down to the formatting of the paper. Influenced by Kaiming He (who even gifted him the Buddhist Diamond Sutra to teach him to look past superficial "forms" to find true substance), Xie believes good taste means avoiding crowded, hyped areas in favor of fundamental, eternal problems like representation learning [02:45:03]. 10. The Degradation of Open Academic AI Research Xie expresses concern over how major industrial AI labs have become increasingly closed off, shifting from open academic exploration to secretive commercial competition. This shift strips researchers of their autonomy and turns them into easily replaceable cogs in a massive engineering machine, heavily motivating his decision to start a new, more open research-driven company [05:21:26].
3
40
3
455
73,385
I packaged up the "autoresearch" project into a new self-contained minimal repo if people would like to play over the weekend. It's basically nanochat LLM training core stripped down to a single-GPU, one file version of ~630 lines of code, then: - the human iterates on the prompt (.md) - the AI agent iterates on the training code (.py) The goal is to engineer your agents to make the fastest research progress indefinitely and without any of your own involvement. In the image, every dot is a complete LLM training run that lasts exactly 5 minutes. The agent works in an autonomous loop on a git feature branch and accumulates git commits to the training script as it finds better settings (of lower validation loss by the end) of the neural network architecture, the optimizer, all the hyperparameters, etc. You can imagine comparing the research progress of different prompts, different agents, etc. github.com/karpathy/autorese… Part code, part sci-fi, and a pinch of psychosis :)
1,042
3,572
1,136
28,107
11,186,331
Very interested in what the coming era of highly bespoke software might look like. Example from this morning - I've become a bit loosy goosy with my cardio recently so I decided to do a more srs, regimented experiment to try to lower my Resting Heart Rate from 50 -> 45, over experiment duration of 8 weeks. The primary way to do this is to aspire to a certain sum total minute goals in Zone 2 cardio and 1 HIIT/week. 1 hour later I vibe coded this super custom dashboard for this very specific experiment that shows me how I'm tracking. Claude had to reverse engineer the Woodway treadmill cloud API to pull raw data, process, filter, debug it and create a web UI frontend to track the experiment. It wasn't a fully smooth experience and I had to notice and ask to fix bugs e.g. it screwed up metric vs. imperial system units and it screwed up on the calendar matching up days to dates etc. But I still feel like the overall direction is clear: 1) There will never be (and shouldn't be) a specific app on the app store for this kind of thing. I shouldn't have to look for, download and use some kind of a "Cardio experiment tracker", when this thing is ~300 lines of code that an LLM agent will give you in seconds. The idea of an "app store" of a long tail of discrete set of apps you choose from feels somehow wrong and outdated when LLM agents can improvise the app on the spot and just for you. 2) Second, the industry has to reconfigure into a set of services of sensors and actuators with agent native ergonomics. My Woodway treadmill is a sensor - it turns physical state into digital knowledge. It shouldn't maintain some human-readable frontend and my LLM agent shouldn't have to reverse engineer it, it should be an API/CLI easily usable by my agent. I'm a little bit disappointed (and my timelines are correspondingly slower) with how slowly this progression is happening in the industry overall. 99% of products/services still don't have an AI-native CLI yet. 99% of products/services maintain .html/.css docs like I won't immediately look for how to copy paste the whole thing to my agent to get something done. They give you a list of instructions on a webpage to open this or that url and click here or there to do a thing. In 2026. What am I a computer? You do it. Or have my agent do it. So anyway today I am impressed that this random thing took 1 hour (it would have been ~10 hours 2 years ago). But what excites me more is thinking through how this really should have been 1 minute tops. What has to be in place so that it would be 1 minute? So that I could simply say "Hi can you help me track my cardio over the next 8 weeks", and after a very brief Q&A the app would be up. The AI would already have a lot personal context, it would gather the extra needed data, it would reference and search related skill libraries, and maintain all my little apps/automations. TLDR the "app store" of a set of discrete apps that you choose from is an increasingly outdated concept all by itself. The future are services of AI-native sensors & actuators orchestrated via LLM glue into highly custom, ephemeral apps. It's just not here yet.
896
998
433
11,959
1,979,392
When I first heard Shinzen and other teachers talk about recognizing depth of practice from the way someone stands, walks, and talks, I didn't know quite what to make of it. Over time, more clearly seeing how this mind is more an extension of these bodily senses than anything apart, noticing ever subtler causal connection between breath, body, and every mind-movement, all these expansions and contractions in awareness, and beginning to recognize the same in others, it made more sense. Then, having countless experiences in-person of my own teachers clearly seeing my own obstructions better than I could, empty mirrors reflecting my own grasping back at me, I had no doubt. Much of what I naively interpreted as metaphor is quite literal.
5
4
2
156
7,817
Joseph Campbell, interviewed by Bill Moyers, on the challenges faced in following one's dreams.
6
97
5
579
28,094
I saw the greatest minds of my generation waste 1000 hours fixating on their breath with no insights or lasting effects bc they read the mind illuminated
67
32
34
1,353
239,491
Joseph Campbell, goddamn
59
519
44
3,459
240,805
Women founders have been reaching out to me over the past 24 hours about how they don’t have permission to run their companies in Founder Mode the same way men can. This needs to change
402
334
247
5,146
1,597,232
We're building a supersonic airliner—powered by the all new @boomsupersonic Symphony engine. Ask me anything!
105
62
11
904
122,618
ai and awakening are possibly the two most highly rated things and i think they're both underrated
8
8
3
173
12,191
Every time I come to Palo Alto, I wonder why I live in San Francisco
49
5
5
326
93,358
this morning i heard @OortCloudAtlas reframe sadness in a way that loosened a knot i didn't know i had in my chest. he said, "usually we're sad about something difficult that happened but if we look behind that, it's because we wish something good had happened it's rly a wish that good things happen; that people and animals feel good it's rly simple in a way, sadness is one of our most beautiful treasures because it's our compassionate heart expressing itself the energy of kindness, caring and love [has] a certain close tie with sorrow"
11
41
3
507
48,099
What a year 2024 has been. Today, we're excited to announce our $6M Series A led by Adjacent to accelerate our platform efforts (Web, Android, iPad) and make @CopilotMoney available to everyone: copilot.money/series-a
48
23
18
360
82,636
Starship will take humanity to Mars
23,262
31,692
4,363
449,573
76,451,692
Congrats to @SpaceX on a successful test flight! Starship has soared into the heavens. Together, we are making great strides through Artemis to return humanity to the Moon—then look onward to Mars.
Liftoff of Starship!
157
1,032
71
10,379
955,628