building the agent computer @...
Joined January 2020
- Tweets2K
- Following819
- Followers983
- Likes3.9K
~2 months since I started building my VMM, now:
- Supports full ubuntu and has cold starts in ~450ms
- Can boot 64VMs, run a simple hello world program and destroy in 366ms
- Next: temporal-like devex for workflow checkpointing, recovery and replay + gpu support
Finishing a few PRs and I will have extremely performant and reliable snapshots, forks and compute relocation
If all you do is make a poor copy of something, don't wonder why you fail.
Being as good as the best is the bare minimum.
My sandbox agrees
AI agents don't use the cloud like humans do. AWS CEO Matt Garman on rewiring AWS for its newest "users":
"Increasingly we're seeing teams that want a cloud that is great to work with agents and not just with people."
" We've doubled down on things like starting a database in three seconds and some capabilities that agents are excited about."
"Many agents wanna create a database, do a little bit of work, and then have the database go away. You don't really need that database to have five nines of durability."
"There are new building blocks that agents are gonna want that we just didn't really need before: compute sandboxes, gateways, agent permissions versus people or service role permissions."
" A lot of those are things that we have built and are building and thinking actively about because they are just brand-new building blocks. It's not using the existing building blocks differently."
@mattsgarman @RaghuRaghuram
Just think higher
Mark Cuban's take on the next job wave deserves more attention than it's getting.
His argument: software as we know it is dying. Generic SaaS gets replaced by AI customized to how each business actually operates.
The question then becomes — who builds and deploys that for the 33 million companies in the US that can't do it themselves?
That's the market. Not AI for enterprises with 500-person eng teams. The other 99%.
SMBs don't have CTOs. They don't have prompt engineers. They need someone to sit across the table, understand their workflow, and wire up an AI system that fits it.
That's a services business. Probably the largest one this decade.
The 6.1-sol was by far the worst rollout the've done in a while
I had to reconfigure all my astra agents because changing the model somehow corrupted them :)
We were not the absolute masters of our universe; we lived in a world that was complex, confusing, and largely uncontrollable.
M. Mitchell Waldrop
Nerd on this topic here! Some thoughts/facts:
- Now very few spanish people has the same socioeconomic conditions to have match what made us become the first global empire. Why?
- We have 3 foundational stages that made us the way we were, the Celtic-Iberian (soul and genetics), the roman empire as the unifier (structure and language) and the Visigoths (blueprint for unified nation).
This 3 eras are a necessary condition to the formation of Spain. (funny enough most spanish people reject this influence, making them arguably worse)
- What made us an empire, made us lose against the British.
Our spirit of independence and local pride made it imposible for the 13 colonies to form a unified, disciplined military command. (We lost, thats why I'm speaking english right now lol)
- We are the only society in history to achieve "reconquista" and we (the spanish) are the ONLY CAPABLE SOCIETY TO DO IT AGAIN (hour long topic + now low iq people thinks it's politics, is not, just history)
Personal opinion on what's EU now: It's the museum of what it used to be. I won't be able to change, so for me... not battle worth fighting.
You can only delay the inevitable
building jio-ubuntu-4v-8g, id be surprised if i'm not the fastest
New hpc-sandbox-benchmarks results are up after running 700+ instances!
We’ve added 2 providers:
1. @boatdotdev
2. tama from @beam_cloud
starsling.dev/hpc-sandbox-be…
Even with my network bottleneck (1Gbps) I'm within 5% of the fastest sandbox provider and I am:
x1.27 faster than sfcompute
x1.97 faster than Daytona
x2.15 faster than Vercel
x2.20 faster than E2B
x3.19 faster than Modal
Now time to scale and time add elastic compute/memory :)
