Mike (SystemSculpt)

Raised by RuneScape private servers and terminal windows. Now I build systems. If there's a messy / impossible process underneath it, know that I'm interested.

[Mike (SystemSculpt) @MichaelStolarz]

you are no longer the violinist in an orchestra,

you are the conductor

[Andrej Karpathy @karpathy]

Expectation: the age of the IDE is over
Reality: we’re going to need a bigger IDE
(imo).

It just looks very different because humans now move upwards and program at a higher level - the basic unit of interest is not one file but one agent. It’s still programming.

[Mike (SystemSculpt) @MichaelStolarz]

@thesaraharminta The second one is a funny joke 😂

[Mike (SystemSculpt) @MichaelStolarz]

@thekitze how do you have access to fable 5 rn

[Matt @ThinkWiselyMatt]

O m g

[Nico Lecomte @NicolasLec91112]

.@CodSpeedHQ uses Claude Managed Agents + @blaxelAI sandboxes to build an AI agent that:

-> ingests customer code
-> benchmarks it
-> iterates on changes
-> finds performance improvements

https://t.co/YCS7q759VE

[Nico Lecomte @NicolasLec91112]

Blaxel is officially a self-hosting partner for Claude Managed Agents sandboxes!

Run the harness on @ClaudeAI managed infra, while task executions run in @BlaxelAI sandboxes for ridiculously fast boot times and full standby persistence.

https://t.co/YCS7q759VE

[Blaxel @blaxelAI]

@ClaudeDevs @e2b @googlecloud @namespacelabs @superserve_ai The best of both worlds: sandboxes that boot in 25ms, hooked right into Claude Managed Agents 🙌 Fantastic to partner with the Claude team to make this happen!

[Kevin Whinnery @kevinwhinnery]

.@blaxelAI sandboxes are persistent and can be resumed from standby quickly (~25ms)

https://t.co/hgjyP1mw57

[Lance Martin @RLanceMartin]

2/ @codspeedhq (a software performance platform) uses Claude Managed Agents w/ @blaxelAI self-hosted sandboxes.

customer code remains within their security boundary w/in the @blaxelAI sandbox.

https://t.co/Dg2GMgWrMe

[CJ Avilla @cjav_dev]

.@blaxelAI's tutorial for managed agents + blaxel sandboxes:

Push orchestrator + worker images to Blaxel, wire a `session.status_run_started` webhook, and every tool call (bash, files, code) executes in an isolated, TTL'd Blaxel sandbox. Secrets stay out via proxy injection.

25ms resumes! ⚡️

https://t.co/k4O8ec52lQ

[Kevin Whinnery @kevinwhinnery]

There's a broad range of sandbox providers for your agent harness with different trade-offs worth considering - highlighting a few that work well with Claude Managed Agents 🧵

[CJ Avilla @cjav_dev]

I wanted to share some Claude Managed Agent tutorials for self-hosted sandboxen to help you get started faster.

It’s pretty sweet how you can have managed agents run code and use tools wherever you want to call “home” as your self-hosted sandbox. (I have so many demo ideas!!)

Here’s a growing list of tutorials that I’ll pile onto as more are built:

[Lance Martin @RLanceMartin]

1/ @RogoAI (a platform for financial institutions) used Claude Managed Agents w/ 10-15k concurrent @e2b sandboxes.

case study w/ video:
https://t.co/J2O2tKUyT4

@e2b docs:
https://t.co/pyyKFP5KD0

[Blaxel @blaxelAI]

Claude Managed Agents is now available on Blaxel.

Run your agent loop on @ClaudeAI managed infra, while running tool execution on secure, persistent sandboxes.

-> 25ms suspend/resume. Shared persistent filesystem. Controllable network.

https://t.co/f0WAQtGjF3

[Paul Sinaï @paul_s3i]

Blaxel can now be used to self-host your Claude Managed Agents sandboxes.

@ClaudeAI runs the agent, @BlaxelAI runs the agent's tasks.

They're two different infrastructure problems. So we split them.

[ClaudeDevs @ClaudeDevs]

Claude Managed Agents can operate in a sandbox you control, on your own infrastructure or with any provider you choose.

Today we added new guides for @blaxelAI, @e2b, @googlecloud, @namespacelabs, and @superserve_ai, so you can choose the best fit for your use case.

[bayes @bayeslord]

The slop era might be coming to an end, to be replaced with an era of endless beauty

[Ethan Mollick @emollick]

"Switch to a cheaper model to save money" is a problem because cheaper models are worse (maybe they are good enough for a particular purpose, but still worse).

More often a better approach is hierarchies of models, with smart models are orchestrators and auditors of cheap ones.

[Ethan Mollick @emollick]

More true than ever.

[shadcn @shadcn]

You have Claude Fable for only a few days. Here's how to make the most of it.

Introducing /improve: use your most capable model to audit your codebase and write plans for cheaper models to execute later.

Studies your code, figures out bugs, perf, tech debt, missing tests, what to build and writes plans any agent can run.

[Mastercard @Mastercard]

As AI agents begin to act, payments move into the background — at machine speed and massive scale.

Today we’re introducing Mastercard Agent Pay for Machines — bringing structure, governance, and trust to this new class of payments.

Launching with 30+ partners to bring this to life from day one.

This isn’t just more payments. It’s a new operating model for commerce.

👉 Learn more: https://t.co/TeS6Lj8jLO