
Not only am I back to agents, slate is getting a massive overhaul
What do you wish your agents could do on their own? And what have you hacked together for them?
Lots of people talking about their software factories
And how much of that swarm scaling is coordination overhead?
Also, the labs are telling you this but you don't need 1000 subagents for a task that only demands 2.
Maybe worth exploring scaling agent count and figuring out where the compute gets spent
I'm finally working on actual agent engineering again.
A few observations:
Seems like everyone has finally caught onto the idea of sliding window compaction.
...
Been messing with Opus 5.5 and boy does this model love to overcomplicate the abstractions it is implementing.
So many things can just be written inline very simply
What's interesting is it is reasoning more clearly and directly about patterns it uses
Pretty good model
I am going to be the Sam Lambert of sandbox CEOs and @freestyle_dev will be the Planetscale.
akira retweeted
Another casualty to sleep deprivation is that I forgot to add the final thanks to this thread...
This is the result of a tons of work from a great team besides myself: @GerardMTFB @akseljoonas @depthfirstsearc @chetantekur
And if you could not tell already the architecture of this harness is inspired by @pidotdev by the goat himself @badlogicgames and OpenAI Codex
Ok can finally talk about what I’ve been working on at Reflection. Introducing Mirror, our own open source coding agent optimized for Beam.
I started the mirror project when I first joined Reflection, and it has gone through many iterations since then.
Install it today with: uv tool install reflection-mirror
github.com/reflection-oss/mi…
akira retweeted
Ok can finally talk about what I’ve been working on at Reflection. Introducing Mirror, our own open source coding agent optimized for Beam.
I started the mirror project when I first joined Reflection, and it has gone through many iterations since then.
Install it today with: uv tool install reflection-mirror
github.com/reflection-oss/mi…
An agent can do all the non-structural parts of systems eng as far as I can tell
This is actually good because you can design systems and think through them and then magically they exist
I feel that certain sandbox providers may have never tried EC2.
t4g.nano costs $0.0042/hour. Worth considering a warm pool of that.
extremely excited to announce our $30m raise for Catalyst.
we are looking for incredible engineers to explore the future of finance as we enter the dawn of superintelligence
please reach out if this is you!
Introducing Catalyst.
What do you believe? How would you trade on it?
You notice data centers are popping-up everywhere, gold will skyrocket, or just know that Labubus will go out of fashion.
Catalyst researches your ideas, builds the optimal strategy, monitors events, and executes trades. Autonomously.
akira retweeted
Everyone in AI is building for everyone else in AI.
We're building for your group chat.
Introducing Charming ✨ Personal apps for us normies.
Make an app for your book club, a meal planner for your family, go wild (it's free!)
Build in Claude/ChatGPT, share like a Google Doc
A prediction:
For the past 9 months there has been a reasonable expectation that every engineer you speak to has a $200 subscription to either OpenAI or Anthropic.
That may be coming to an end with the rise in open models. Companies like baseten may offer inference passes and companies like Devin and Random Labs may become more serious contenders.
This hinges on cost per task for most tasks agents complete continuing to drop at current rates.
As models get smarter I am more and more convinced that @a1zhang's mismanaged genius nomenclature is representative of whats going on
There is no reason for these models to struggle the way they do at this point









