education @simdotai, AI R&D @IdyllicLabs, prev: dev rels @terra_money, head of R&D for wasm devx @cosmology_tech & @terran_one, @ucberkeley ('19 dropout)

Joined October 2015
862 Photos and videos
Pinned Tweet

4
3
23
7,123
Will Chen retweeted

150
1,556
9,381
3,279,938
Will Chen retweeted
Introducing Mirage, a unified virtual filesystem for AI agents! 6 weeks. 1.1M lines of code. We rewrote bash from the ground up so cat, grep, head, and pipes work across heterogeneous services. S3, Google Drive, Slack, Gmail, GitHub, Linear, Notion, Postgres, MongoDB, SSH, and more, all mounted side-by-side as one filesystem. Bash that AI agents already know works on every format! cat, grep, head, and wc parse .parquet, .csv, .json, .h5, even .wav! One pipe can stitch S3, Drive, GitHub, Slack, and Linear together, same Unix semantics throughout. Workspaces are versioned too. Snapshot, clone, and roll back the whole thing with one API call. A two-layer cache turns repeated reads into local lookups, so agent loops stay fast and cheap. Drop a Workspace into FastAPI, Express, or a browser app. Wire it into OpenAI Agents SDK, Vercel AI SDK, LangChain, Mastra, or Pi. Run it alongside Claude Code and Codex. Site: strukto.ai/mirage GitHub: github.com/strukto-ai/mirage #AIAgents #OpenSource #AgenticAI #Strukto #Filesystem #VFS
171
338
3,269
618,991
youtube shorts is our generation's smoking
1
10
845
default creative mode is ergodicity
4
483
a company is a human harness
3
563
certain formulations arrive with unusual authority they feel found, not manufactured taste seems like tuning, not arbitrary preference repeated exposure refines receptivity creation feels partly receptive
1
5
545
people mistake rawness / rough for authenticity the most authentic voices feel exceptionally eloquent and articulate at random moments many writers report perfectly phrased sentences pop out of nowhere, from the spirit world and build the essay / paragraphs around those sparks
5
554
what is missing in modern life that causes so much drift, self-doubt, and non-compounding effort?
1
552
explains the confusing workers ecosystem from first principles
Durable Objects Facets were just added to @Cloudflare for Agents Week There's now four ways of running AI-generated code on this AI cloud → Dynamic Workers → Sandbox / Container → Durable Object Facet → Workers for Platforms Detailed breakdown below 🧵
4
970
share your golden orb / alpha signal
Apr 10
Here’s a cheat code: find something that fascinates you and then share your fascination. The energy of genuine fascination is the most compelling force in human communication. More than eloquence. More than authority. More than charisma. A person who is visibly, unreservedly excited about something is magnetic.
1
3
668
matches with my intuitive sense of "agency" -- agency = your ability to be CAUSAL, you possess it when you feel like you can make shit happen ∴ high agency = high causal leverage
literally yes agency does not = working hard high agency = accomplishing your goals with as little effort as possible, since there are endless good uses of your time and energy, and spending less of them where they’re not needed frees them up to spend somewhere else
1
5
1,593
I have coding agents try to do 100 different experiments in 100 different dimensions, search my past, search the web, search the last four months back, test with some skills for simulating myself, prepare thoughts I probably would think of. Then that set of 100 different experiments I can index, I can embed that, that becomes this new ensemble that has statistical properties that are interesting. You can see the central limit theorem applied to yourself.
1
2
523
Will Chen retweeted
We built Missions at Factory, and I wrote about the architecture that I led the design for to make multi-day autonomous coding reliable. Agents are highly reactive to their context. Every design decision follows from keeping each agent's trajectory focused and directionally consistent.
15
32
341
30,194
i call this "ensemble engineering" as using multiple agents becomes common, the only way to say bitter-lesson-proof is to generate like 300 different strategies all running loops in parallel and sample from that distribution
2
707
cardio, 35min with thunderstorm sounds, daily
Doomscrolling destroys the one thing ancient cultures considered most sacred: the integrity of your inner temple. Every wisdom tradition understood that consciousness wasn’t just awareness. It was a consecrated space that required careful curation of what was allowed to enter. The Greeks called this “mneme” - not just memory, but the deliberate cultivation of what deserves to be remembered. Roman orators spent years training their minds like architects designing sacred buildings. They created “memory palaces” - elaborate mental structures where only the most important knowledge was permitted residence. Every piece of information had to earn its place through repeated contemplation and integration with existing wisdom. They understood something we’ve forgotten: your consciousness becomes identical to what you repeatedly allow into it. Doomscrolling inverts this ancient principle entirely. Instead of carefully selecting what deserves mental residence, you open your consciousness to a fire hose of random information that demands attention but deserves none. Ancient philosophers would be horrified not by the content of what you’re consuming, but by the complete abandonment of discernment about what belongs in your mind. Marcus Aurelius wrote his Meditations as daily exercises in consciousness curation. Each entry was him deliberately choosing which thoughts deserved his attention and which should be dismissed as “vain imaginings.” He treated his mind like a fortress that required constant vigilance about what was permitted entry. Doomscrolling is the opposite: throwing open the gates and inviting every passing thought, emotion, and reaction to take up permanent residence. Buddhist monks spent lifetimes mastering “mindfulness” - not the watered-down modern version, but the rigorous practice of watching what arises in consciousness and choosing what to feed with continued attention. They knew that whatever you attend to grows stronger. Whatever you ignore eventually fades. Doomscrolling feeds everything equally. You give the same quality of attention to genuine crises and manufactured outrage, to problems you can solve and problems that don’t actually exist, to information that improves your life and information that only increases your suffering. Ancient wisdom keepers understood that consciousness without curation becomes chaos. They developed elaborate practices - meditation, prayer, philosophical dialogue, contemplative reading - specifically to maintain the sacred order of their inner world. We’ve replaced these practices with aalgorthmic feeds designed by people who profit from your psychological disorder. The Sufis had a teaching: “Your heart is a polished mirror. You must wipe it clean of the dust that has settled upon it, because it is destined to reflect the light of divine secrets.” Doomscrolling doesn’t just add dust to the mirror. It convinces you that dust is what mirrors are supposed to reflect. Ancient cultures treated consciousness like a garden that required constant tending. They knew that if you don’t deliberately plant what you want to grow, weeds will take over everything. Your consciousness has become a wilderness of weeds that you mistake for staying informed. The ancients were right. Your inner world is sacred space. What you allow to dwell there shapes not just how you think, but who you become. Doomscrolling desecrates that space by treating it like a garbage dump for every piece of information that demands attention. Your ancestors would weep.
3
482
today, i declare independence from nicotine, youtube, and doordash @japarjam
2
8
816
Need to mute the word “introducing” on this platform
1
4
377
Will Chen retweeted
don't build agents at all, build in real life relationships Trust me
instead of building web apps build desktop apps trust me
2
2
24
2,452
Will Chen retweeted

5
1
18
1,279