
agents @LuxorCloud bitcoin @Luxor hashrate @HashrateIndex
Earth
Joined March 2012
- Tweets3.1K
- Following3K
- Followers1.6K
- Likes4.5K
what's tenki?
tenki is now luxor agent cloud (@luxorcloud) → luxor.ai
fun fact: tenki means weather in japanese. turns out we were a cloud company all along.
why luxor? tenki was always built at @luxor. same team, same metal, now with all of luxor behind it.
what we’ve shipped:
- runners, sandboxes and code review on full linux VMs, on servers we own
- @tailscale native across CI and sandboxes
- #1 sandbox provider by @lightsage CLI agent usage on devtool arena
- live in @Cursor_ai Origin
- plug into @tinyfish, @hotdata, @omnara, @naive and @anchorbrowser so agents get browsers, databases and web tools out of the box
- per-second billing, no rounding tax
- Ryzen 5 9950X CPUs, NVME CEPH, 400 gbps back plane
the growth over the last few months has been unlike anything we’ve seen. agent cloud is luxor’s future, so we’re going all in: a lot more hardware, dedicated sales, and the full 110 person luxor team behind it.
what changes for you? nothing breaks. tenki.cloud redirects and your setup keeps working.
agents write, test and review the code. we run all of it.
1
6
178
eddie 王 retweeted
Our most prolific engineer at Lygos doesn't write any code. She just reads all of ours.
Tag her in a PR at 3am on Saturday? PR is reviewed before we even wake up from our nap💤
@tenki has reviewed 600+ of our PRs since June.
So excited to announce that we're bringing @Tailscale to sandboxes with native network isolation that even the sandbox's root user can't disable.
Tenki sandboxes can now join your @Tailscale tailnet. Your ACLs decide what an agent can reach. Ephemeral sandboxes leave your tailnet when they shut down.
5
216
eddie 王 retweeted
New: @Tailscale connections for tenki.cloud runners and sandboxes 🔒
- Connect your tailnet once: add a Tailscale connection under Settings → Networking. It uses OpenID Connect, so Tenki stores no Tailscale auth key or OAuth secret.
-GitHub runners: make the connection the runner default for chosen repositories, or create a dedicated tenki-tailnet-* runner label. Jobs on that label join your tailnet before the first step, with no Tailscale step or GitHub secret in the workflow.
- Job access controls: limit tailnet access to protected branches, same-repository jobs, or all jobs including fork PRs.
- Sandboxes: set the connection as the default for every new sandbox, or attach it to one sandbox with tenki sandbox create --tailnet-profile-id <connection-id>. Auth keys still work if you prefer them.
- Built-in connection test: check that the connection joins your tailnet and can reach a private endpoint before you route CI through it.
📖 Docs:
- Runners and saved connections: tenki.cloud/docs/runners/tai…
- Sandboxes: tenki.cloud/docs/sandbox/tai…
@tenki is building the true enterprise cloud for agents
eddie 王 retweeted
tenki.cloud is the fastest and affordable CI provider
be like Leo, use Tenki
I had my agent benchmark this against my current CI provider (@tenki)
At the moment choosing Vercel CI doesn't make sense as it's slower in most cases, and more expensive.
But it will only get better from here! 🫡
eddie 王 retweeted
new hardware is online
benchmarking ~30-40% faster than our previous fast machines.
and we killed premium pricing. everyone gets the faster machines at standard rates, no upgrade needed.
net result: half the cost of GitHub-hosted runners while being 3-4x faster.
tenki.cloud just got cheaper and faster.
Saturdays are for racking and stacking more compute nodes for tenki.cloud, these high density microclouds are to increase our capacity on our CI runners
Effective Immediately, all organizations using PHP for their systems are advised to be prepared for an update addressing two critical vulnerabilities. PHP is widely used for web applications. These vulnerabilities often provide a direct path to the server's file system or database. If an immediate update is not feasible, it is recommend that organizations take the system offline until it can be fully mitigated. NSA will follow up shortly with more specific guidance.
eddie 王 retweeted
tenki.cloud is one of the first native apps on Cursor Origin. AI code review is live today. CI runners are next.
huge thanks to the @cursor_ai team for helping us ship it.
eddie 王 retweeted
Origin, Cursor’s code hosting platform, now supports over 20 integrations.
Connect your repos to tools for CI, deployments, code review, security, monitoring, and project management.
eddie 王 retweeted
new compute is going live at tenki.cloud, so we're retiring our slower machines. every runner will run on fast hardware at the normal rate. no more premium pricing.
.@OpenRouter co-founder Alex Atallah on the "everyone is building the same thing" take:
"This reminds me of this tweet I saw... Everybody's building an agent loop with notifications, third-party connectors, context management, memory, sandboxes, agentic web search, an always-on agent on top."
"This product is showing up everywhere. It is showing up everywhere, but these are just the new table-stakes primitives."
"A 2005 version of that tweet would be: 'Oh, everybody's building the same thing. A database, a users table, a sign-in page, a sign-up page, a profile page, a logout page. Everything's the same.'"
"There's like a lot of differentiation really. There's table-stakes needs for AI just like there's table-stakes needs for the web."
@alexatallah @amasad
.@OpenRouter co-founder Alex Atallah, in his first podcast since Stripe acquired the company, joins @Replit co-founder Amjad Masad and a16z's Erik Torenberg on why the future of AI is independence and specialization.
In this conversation, Alex walks through how the Stripe deal unfolded, why he wasn't originally looking to sell, and why "payments and inference are going to blend together."
Pre-OpenRouter, the typical AI workflow had one model provider to choose from, and little pressure on that provider to lower prices. Now enterprises are diversifying across labs and open-weight models, and every board is asking about AI costs and benchmarks.
Amjad argues if your company depends on one AI lab, it can turn into your competitor. So Replit is building the layer that lets enterprises use any model and any cloud, without being locked into either.
Alex and Amjad are split on personal agents – Amjad runs one agent across his whole company and loves the cross-domain joins, while Alex says general agents cause you to sacrifice understanding, and argues 10 specialized chiefs of staff beats one superagent.
0:45 How the Stripe deal unfolded
5:05 Why mixing models beats one model
7:25 Forcing the labs to compete on price
8:50 Enterprises want open-weight models
10:30 Every board asks about AI every month
12:25 Why companies must own their intelligence
14:15 Replit as the independence layer
15:10 Everyone is building the same agent
16:35 Why Amjad built bring-your-own-cloud
18:10 Amjad's agent that runs his whole company
19:55 Why 10 specialized agents beat one
23:35 Machines, not humans, should specialize
27:30 Guardrails for agents talking to agents
31:10 Models training their own replacements
33:45 Most tasks don't need a frontier model
40:50 Training small models on Qwen 8B
43:25 The Rust cycle is coming for AI
45:10 Fusion models: frontier quality at half the cost
YouTube: youtu.be/ekK8urKHPMQ
@alexatallah @OpenRouter @amasad @eriktorenberg
eddie 王 retweeted
new in tenki.cloud sandboxes: secrets that never enter the VM.
agent code uses a placeholder, and the host swaps in the real token on outbound requests to allowlisted APIs. nothing in the sandbox to leak or exfiltrate.
tenki.cloud/blog/sandbox-sec…
eddie 王 retweeted
Agents get blocked at the browser. Sandboxes give them a machine but they still can't get into the sites they need.
tenki/anchor:stable is our first pre-built template. Standard sandbox image + playwright-core, connected to an @AnchorBrowser managed browser over CDP. The API key is passed per session and never lands in the image or a snapshot.
Run: `tenki sandbox create --image tenki/anchor:stable`
Docs: tenki.cloud/docs/sandbox/anc…
Yeah, this is exactly what I've been telling people since February.
sluongng.substack.com/i/1853…
But I know Anthropic hoarded a bunch of Bazel/Buck/Nix talent in their ranks, so it's just a matter of prioritizing the right problems at the right time.
At Anthropic, Claude now writes 80% of our code. Engineers ship 8x more code per quarter.
Side effect: Tests grew 10x. CI jobs up 25x in 6 months. Here's what helped us scale:
claude.com/blog/agentic-codi…
eddie 王 retweeted
Kicked off the Startup Program in SF.
100+ builders in the room.
Thank you to everyone who came out to build with @tenki.
eddie 王 retweeted
We were glad to support the launch of @nerdconf_ar's Burning Token hackathon this weekend! The event is going all week: burningtoken.dev
this is the way 🫡
I have lamented this
I have come to peace with this
I prefer optimism
Here is a live view into omarchy while an agent navigates to the lock screen, with the model intent and the series of actions completed on the side
I love this, even if I am slopping it. Making something developers will, hopefully, use and make their life better is literally what got me excited to begin with. I am still doing it, going to keep doing it.
3
275
eddie 王 retweeted
some photos from friday's @tenki for startups launch
100+ engineers and founders came through. good conversations about agents, infra, and what people are actually building




























