Daniel Worku

Co-founder @starslingdev 💫 Previously NFLX, META and many startups

[Daniel Worku @dbworku]

@Caseymanos Great analysis on the nitro sandboxes! Appreciate the usage of our benchmarks

[Daniel Worku @dbworku]

@ryan__yao Congrats! You all are doing great.

[Daniel Worku @dbworku]

@CalvinGrunewald Hot take. Most companies waste CI compute. Hardware doesn't solve that.

I'd be curious what our /ci-speedup skill finds, If you have some spare tokens
https://t.co/zT32EAxssX

[Daniel Worku @dbworku]

@AniC_dev I want to see Boat succeed and utilize your service. But keep it honest.

There’s a major gap between the current Boat quality / performance for larger images. Cheap yes but not fast and reliable for critical infra.

[Daniel Worku @dbworku]

@JohnPhamous this is good. thanks!

[Daniel Worku @dbworku]

@nikitabier Why is time to first still seconds? And why is slower than before. We expect faster from other social media.

[Daniel Worku @dbworku]

Pro tip: for smaller DBs go no disk.

Tmpfs on RAM. Much faster than even NVMe

[Ben Dicken @BenjDicken]

TL;DR your database should be on local NVMe

[Daniel Worku @dbworku]

@jonobelotti_IO Exactly. Kudos for Modal for being one of first with active usage based billing.

[Daniel Worku @dbworku]

@bradgessler If you want high core count, I think we are one of the few with 64 core x86 runners.

[Daniel Worku @dbworku]

@altcap @TrumpAccounts @InvestAmerica24 Trick or treat Brad!

[Daniel Worku @dbworku]

Lie: "CI is hardware constrained"

The top tech companies in the world are solving the great CI bottleneck of 2026 with code changes. You can too with our agent-friendly playbooks.

[Yonas Beshawred @yonasbe]

A new GitHub Actions runner launches every week now. But getting faster runners is just one part of what the best engineering teams do to speed up their CI.

Introducing CI Playbooks ▶️ an agent-friendly library of 253 GitHub Actions optimizations from companies running software factories.

We analyzed 416 pull requests and commits, and 182 blog posts, talks, and podcasts from ten companies:

1. Anthropic: CI job volume grew 25x in six months as Claude came to write 80% of its code

2. DoorDash: build and test cut from 22 minutes to 6

3. Linear: jobs 34% faster while test suites almost quadrupled

4. Mastra: 29m 56s to 5m 06s p95 on StarSling runners

5. OpenAI: about 30 CI PRs took median CI from about 10 minutes to 4.5

6. Ramp: its coding agent accounts for 74% of merged PRs

7. Shopify: an agent running experiments made the Polaris build 65% faster

8. Spotify: deployer image build went from 2h28m to about 17 minutes

9. Stripe: runs about 5% of its full Ruby test suite per build

10. Uber: CI resource usage down 53%, wait times down 37%

Some of these companies aren't using GitHub Actions as their CI engine internally so we translated all the optimizations they've talked about or made publicly in their repos, into their GHA equivalents.

Every technique comes with its original source and a prompt your coding agent can run on your repo. Add .md to any playbook URL for the markdown version, or just hand this URL to your agent:

https://t.co/dngjnfBb99

There's tons to dig into here, I'm going to highlight some of the findings I found most interesting 🧵

[Daniel Worku @dbworku]

@jarredsumner Crazy idea Jarred...

AOT on a subset of JS -> to continuous mathematical equations

Basically what if Bun could output a jax, taichi lang or handwired style differentiable model for a GPU out the other side of compile

[Daniel Worku @dbworku]

The control plane piece of Vercel CI is good.

Fast booting VMs that execute each task in the optimal hardware environment concurrently.

[Vercel @vercel]

Introducing Vercel CI

• Intelligently skips tasks with Turborepo
• 2-6× faster CI in our internal repositories
• Optimized for agents via Vercel MCP and CLI

After nearly 8 million tasks, we're excited for the future of agentic CI. In beta on all plans.
https://t.co/IcsuvweAhB

[Daniel Worku @dbworku]

@dominikkoch @techpoto @nikunjatrey @starslingdev We offer 2k minutes free but happy to enable one free agent PR.

DMs are open

[Daniel Worku @dbworku]

@arthur_guiot @useblacksmith 😢 it does not need to be this way.

A multi-geo and multi-sandbox architecture routes around precisely failure with traditional CI providers.

[Daniel Worku @dbworku]

Concentrating high-performance hardware in a few datacenters creates a shared failure point.

A regional power or cooling failure can take every runner hosted there offline.

CI needs speed and a resilient router architecture.

[Daniel Worku @dbworku]

@techpoto @nikunjatrey @dominikkoch @starslingdev Over time, agent PRs optimize time and reduce usage with the same techniques that Anthropic, OpenAI and Linear engineers use.

https://t.co/0rbuDejaw8

[Daniel Worku @dbworku]

@techpoto @nikunjatrey @dominikkoch @starslingdev Immediately, adaptive compute enables competitive real-world performance that is more resilient than single provider CI.

https://t.co/AYrHkzATOl

[Daniel Worku @dbworku]

@jarredsumner 👀 @rajashree