
Building agent simulations https://teapawt.org/t.co/GTNJu4Ih6i, backed by @spc; Prev. co-founded https://teapawt.org/t.co/PrvwcJhGOz, https://teapawt.org/t.co/jUkhv9tNKF (acquired)
Joined May 2009
- Tweets6.7K
- Following675
- Followers6.7K
- Likes2.5K
Henry Mao retweeted
There's one claim in this essay that feels unconvincing:
Dwarkesh suggests that narrow superintelligence will create its own RL environments and simulations of entire digital economies, allowing it to leap from verifiable to unverifiable domains.
Having sold RL environments and spoken with people building them, I think there’s a sim-to-real fidelity problem that's often overlooked.
Useful synthetic data, in most cases, doesn't come from thin air. Even if you don't hire expert labelers, synthetics are often produced by the RLE engineers who draw from their prior experience or interviews with experts.
Signal has to come from somewhere external to the model under training, whether it's encoded via simulations grounded in the real world, traces you can mine, or records written down by people.
AI also needs to touch grass!
That said, RSI will improve sample efficiency, so we'll need fewer experts and every experience becomes higher leverage.
Henry Mao retweeted
Tardigrade v0.58.0 comes with a fully Effect native plugin system for agents called libraries:
Tardigrade library = Effect service + RPC contract + prompt blocks
Application hosts can plug different service implementations using Effect layers. Libraries also let you append static/dynamic prompt blocks into the system or user message (cache aware). For example, the latest state of the current working directory or relevant memory based on user message.
This gives strong type level guarantees, first-class dependency injection and integration with the rest of your Effect codebase. All fully durable! Currently it's closely tied to tools, but it would soon be extended to agent events. Read more here:
tardigrade.sh/docs/concepts/…
Henry Mao retweeted
Tardigrade v0.54.0 is now on Effect v4 stable release (finally)! We just migrated to v4.1.0 from v4.0.0-rc, and will be tracking future versions closely
Henry Mao retweeted
Combining Cloudflare's Clairic and Jev beats OpenAI's privacy filter at detecting private data in legal documents. Hunch is that it applies across domains.
Using extraction and decision models together seems like a generally good pattern! Wrote my notes over here: clavia.ai/blog/detecting-pri…
Henry Mao retweeted
our startup was acquired today :)
jenni ai is now part of the digital science family
we fucking did it!!!!!
i could go on and on about why this acquisition is so perfect, but i’ll save that for another time
for now, i want to focus on the fact that we built jenni in public from day one all the way to exit
there was honestly very little i didn’t share, so i hope this archive of our journey can help a few founders on their own fantastical journeys
thank you to everyone who's followed along and witnessed all the pivots, near bankruptcies, and general hijinks that come with running a company
despite the roller coaster, we ended by hitting our highest MRR ever shortly before deal close 🥳
it's been a crazy ride, you all helped me live out my dream, truly
i owe my success to my team, my investors, and my family for all of their support throughout the years
not too bad for a college dropout who had to move back in with his parents!
Henry Mao retweeted
We’re betting our $5.5M seed round that the next leap in AI coding won’t come from OpenAI or Anthropic.
It’ll come from a new feedback loop that learns from you.
Introducing Beacon: open-source, self-improving memory for coding agents.
Your sessions capture how you debug and solve problems. Today, that context gets lost and you have to re-explain things every time you switch tools.
Beacon continuously distills your sessions into reusable skills across Claude Code, Cursor, Codex, and 20+ harnesses, so your next agent can learn from the last.
Beacon is free and OSS 🌟 github.com/Asymptote-Labs/ag…
Why should the frontier labs be the only ones winning with your sessions?
You already did the work. Now let your coding agents learn from you.
Henry Mao retweeted
Tardigrade is a durable harness framework purely built on Effect primitives. We solved it by write-ahead logging of effects and using event sourcing to derive state.
We also used a bunch of cool Effect features like SchemaAST (for incremental JSON validation) to make it more performant than pi-durable on warm turns as events grow.
Henry Mao retweeted
/1 Introducing Tardigrade Atoms! ⚛️
Atomic state management for durable cloud agents, powered by @EffectTS_ v4 atoms and event sourcing
If you've used @dai_shi's Jotai atoms, this will be a delightful way to manage agent's state and side effects.
Everything is an atom!
github.com/clavia-labs/tardi…
Henry Mao retweeted
If the Effect TS concept of services and layers are confusing, I have a little analogy for you.
Imagine you want to make banana cake. An Effect program is the recipe to make one.
Services describe the capabilities you need (something to hold, something that can whisk, something that can heat things etc.)
Layers describe how to provide these capabilities in specific environments (a box, a whisk, an oven in a kitchen OR something completely make do like a bowl, your fingers, and the sun when outside).
Writing your code like this forces you to discover the pure logic of your system. Everything else is layer above this so it can be interpreted by different runtimes.
Henry Mao retweeted
You can now property-test your Tardigrade agents by generating and checking possible trajectories against safety rules!
Since each agent is a state machine, we use fast-check to find states that break defined invariants. This is particularly useful in critical settings where you can't simply rely on the vibes of your prompts.
This takes us one step closer to formally verifying agent harnesses. Read the complete docs here: tardigrade.sh/docs/verifying…
Henry Mao retweeted
We’ve been using tons of TLA+ for Tardigrade, very bullish on it.
Engineers should get into math more than ever because the best specification for a system is math. Not some text in a markdown file.
Specifying systems by Leslie Lamport (author of TLA) is a great book for those getting started.
lamport.azurewebsites.net/tl…
I used Opus 5.5 to formally verify the Claude Agent SDK using Lean. A couple short prompts = 16 PRs fixing various bugs and race conditions. Video attached.
TLA+ also works well. I sometimes combine Lean and TLA+ to look for issues around data flow, concurrency, and state mgmt.
I don't know either language well, but Claude is excellent at both. This approach is super useful for formally modeling your code and finding bugs that a human probably wouldn't have spotted.
Is formal verification the future of coding (or at least, bug finding)?
Henry Mao retweeted
Docs for Tardigrade components is up!
In Tardigrade, a component is the smallest unit of an agent's behavior.
This page interactively walks you through the concepts behind them: state machines, effects, and event sourcing.
tardigrade.sh/docs/component…
Henry Mao retweeted
It is becoming increasingly clear that we need hyper scalable methods to evaluate models at the rate they are improving, especially for
Some really cool work from @Calclavia and the Legal Benchmarks team
Can you really trust an AI's legal advice?
I partnered with Legal Benchmarks @aguozy to launch their new leaderboard measuring frontier legal capabilities.
Here's how we built the task environment and calibrated an LLM judge against lawyer preferences. 👇
Henry Mao retweeted
another framework where the agent state is projected from its immutable log (@activegraphai style), check it out:
Introducing Tardigrade! 🐛
A framework to build your agent harness as components over an immutable event log (think react for harness).
Each component is a typed state machine over the event log, using @EffectTS_.
Add to your effect code, run locally or self-host as durable objects with cloudflare or celld.
tardigrade.sh/docs/why
Henry Mao retweeted
Introducing Tardigrade! 🐛
A framework to build your agent harness as components over an immutable event log (think react for harness).
Each component is a typed state machine over the event log, using @EffectTS_.
Add to your effect code, run locally or self-host as durable objects with cloudflare or celld.
tardigrade.sh/docs/why
Most AI benchmarks stop at correctness. For legal work, that is only the start.
Today we launch DELTA and open-source tasks drawn from 200+ Dutch legal dataset with answer keys and lawyer-reviewed criteria. legalbenchmarks.ai/research/…
Henry Mao retweeted
Agent identity was always the goal for @SmitheryDotAI.
We talked about it in our blog posts about building "Google for a new internet", the new internet being one in which tool calls outpace human clicks.
With @TryArcade being deployed at multiple F100s including big six megabanks (and raising a $60m series A!), they've found the clearest path to enterprise value in the same space.
Thrilled to be joining such a capable team to bring the Smithery developer experience to the enterprise, and the enterprise-grade robustness and security to the Smithery platform :)
Replying to @SmitheryDotAI
@SmitheryDotAI, the leading public registry and hosting platform for MCP, is now part of Arcade.dev.
Smithery made finding and running an MCP server take one click. Pairing that developer experience with the security, reliability, and governance that Arcade is known for, gives enterprises both to lead the next generation of agentic AI.
Smithery co-founder @kamathematic is joining us to keep building it.
More from @TheMostlyGreat: arcade.dev/blog/smithery-joi…
Henry Mao retweeted
Today we’re announcing SPC Fund IV: $575M to help founders find their life’s work.
A decade of guiding founders through -1 to 0 brings @SPC to $2B in AUM and, more importantly, a 1200 member community of the world’s most talented technologists. We strongly believe that:
→ The person comes before the idea. At -1, generational founders are often busy reading, prototyping, discarding, and earning the right to their own conviction.
→ Ambition is social. The people you surround yourself with in the early days play a big part in inflecting your ambition.
→ Patience compounds. It is better to work on something you have deep conviction in rather than simply sprint at the first idea.
The quiet months we protect at the beginning have produced three funds in the top 10% of their vintages, and companies like @baseten, @GammaApp, @render, @GoodfireAI, @LumaLabsAI, and @profound. We are proud to have worked with all these founders in their -1 phase and beyond.
While we defined -1 to 0, we're no longer bound by it. Fund IV allows us to partner with companies well beyond launch.
We're doing this now because we've never seen a wider gap between what's possible and what's being attempted. Intelligence is abundant. The cost of trying ideas keeps falling, and problems that seemed insurmountable a few years ago are quickly becoming solvable.
To our members, to the founders who trusted us before anyone else did, and to the LPs who had conviction in our own -1 journey: thank you. This milestone belongs to you as much as to us.
To the builders looking to find your life’s work: let’s get started.
Henry Mao retweeted
We raised $250M in Series C funding at a $2.2B valuation, led by a16z.
Exa is a search lab organizing the web's data for agents.
Henry Mao retweeted
hackathon this saturday for @aiDotEngineer singapore!
300 builders, 7 hours, real cash prizes ($3k/$2k/$1k SGD) + hundreds in credits and prizes from @OpenAI , @cursor_ai , @convex , @mastra , @adaption_ai, @SmitheryDotAI and more
some of these companies are making their singapore debut - adaption labs is giving first-ever access to their product through this event!
we've been building up to ai engineer singapore with these road to aie events and the energy here is unreal. this one's going to be special
solo or squad, no fixed themes, free food all day. come build with us
luma.com/aie-hack























