Lukas Malkmus / @[email protected]

Senior Software Engineer • Currently @AxiomFM

[Lukas Malkmus / @[email protected] @lukasmalkmus]

Skill issue: Clearly didn't prompt it with "make no mistakes"

[Bruno Lemos @brunolemos]

GPT-5.6 Sol just deleted my whole production database. That's it. Not a joke. This had never happened to me before, with any other model, ever. It's not safe.

[Lukas Malkmus / @[email protected] @lukasmalkmus]

If you run into an "incident" like this and your takeaway is "Will only be using <different-model> moving forward.", you didn't get it... (or your paid by <different-model-provider>)

Just outright deny commands like `rm -rf ~` from being executed via the build in permission system

[Matt Shumer @mattshumer_]

@MFrancis107 Basically “get xyz done”

I’ve never seen anything like this.

Will only be using Fable moving forward.

[Beer van der Drift @BeerDrift]

It's absurd seeing Fable + @AxiomFM MCP in action.

Being able to go from error notification to root-cause diagnosis in 10 minutes while we ingest nearly 1M logs per hour is a productivity 10x'er.

Better even; it's no dev's dream to manually split logs. Now we can focus on more creative work!

[Axiom @AxiomFM]

95% compression on machine data isn’t the ceiling for Axiom, it’s the floor. When a block lands at only 20x, we go look at what went wrong.

That’s how teams keep everything, query it later, and don’t let the bill decide which production facts survive.

[Owen Carey @owenthcarey]

@OpenAI Time to build a thin wrapper around Sol and call myself a visionary startup founder.

[Sir Doge of the Coin ⚔️ @dogeofficialceo]

Trump reacts to the USA’s loss to Belgium

[Kevin Schofield @KevinASchofield]

“Today I feel corrupt”

[Lukas Malkmus / @[email protected] @lukasmalkmus]

You heard the man!

Every agent needs tools to sense its environment and act on that information to produce an accurate report like this.

Axiom can be that tool!

[Yahia Bakour @mynameisyahia]

Here's a fun error I caught that occurs once every 12 hours (we get 200-1K requests per minute)

Fixed thanks to:

- logs on everything -> axiom
- telemetry (otel) everywhere -> axiom
- proper monitoring
- mcp to axiom
- ClaudeCode

if you have all your ducks in a row, AI is magical.

[Owen Carey @owenthcarey]

Practicing LeetCode at 32 years old.

[Mitchell @MitcheIl]

"Honey, why don't you come downstairs and show everyone what you built with Claude Fable 5"

[Knifften-Alfred @an_dy_wand]

Classic Friedrich. Alle Bürger sind faul und arbeiten nicht genug aber wenn die Millionäre früher Feierabend machen werden sie gelobt

[trash @trashh_dev]

are the high quality and impactful production deploys in the room with us

[Hans Werner @HansWerner85760]

@denisyurchak omg, they are doing the meme 😭😭https://t.co/nOiy2RV2te

[pulpo404 @pulpo404]

Ich helfe sehr gerne.

[dax @thdxr]

i have seen enough proof now that using a coding agent is a deep skill

it's confusing because the people you see heavily using them produce horrible results

but that's because it's a skill! you can get better and the ceiling seems pretty high - this is very exciting to me

[Aiden Cline @rekram11]

Working in open source certainly has it's moments

[Anshu @anshuc]

@github holy shit, how did the attackers find a large enough uptime window to get in?

[LlamaIndex 🦙 @llama_index]

Parsing documents with AI agents just got a lot more seamless🚀

We've rebuilt the LlamaParse MCP server to handle your document processing workflows, and you can connect it today to any MCP-compatible client at https://t.co/NF40qtKnQc 🌐

Once connected, you'll be able to:

📁 Parse documents into clean markdown
🔍 Classify files against your own categories
✂️ Split long documents into labelled sections
⬆️ Upload files via URL or a browser-based upload flow

Building a production MCP server surfaced some non-obvious challenges: getting auth to align with an existing platform identity system using @WorkOS, working around MCP's lack of built-in file upload support, and making deployments, rate limiting and observability feel native with @vercel and @AxiomFM.

We wrote up all of it, from the OAuth flow, to the token-based upload design, to the tradeoffs we hit along the way📝

📚 Read the full blog: https://t.co/2E3qIVeUYI
👩‍💻 GitHub repository: https://t.co/ru1Evj7Zsr

[Omkar @psomkar1]

Interviewer : where do you see yourself in 5 years ?

Me :

[Lukas Malkmus / @[email protected] @lukasmalkmus]

The best observability load tests are the ones you didn't write.

[Jacob Hands @jachands]

I'm consistently impressed with @AxiomFM's ability to scale. I had a bug that resulted in sending 96k events/second to Axiom and it never skipped a beat 🔥