Laís Carvalho
@lais_bsc
DevRel & Growth Marketing @Pydantic | markdown dev | PSF & Europython Fellow | PSF Community Service Award 2024
647 Following 853 Followers
Now, you can use gh aw to let your custom @pydantic
AI agent maintain your repo.
Have you tried the GitHub agentic workflows engine?
AWS Lambda durable functions now integrates with @pydantic AI harness
The @youdotcom Answer API verifies every citation exists in the source text before it returns—93.48% on SimpleQA, p50 2.67s. 🔍
@youdotcom is now available as capabilities in the @Pydantic AI harness. See how we wired into a Pydantic AI M&A deal desk: three agents, three research surfaces, one auditable trace.
Show more
Interesting comparison on LLM observability tools.
I haven't heard of Retrace or Prefactor but my heart is full to see Logfire in #2#.
That's my fully unbiased opinion for today.
26 LLM observability tools, measured and ranked.
Field median 77. Best 98. Floor 54.
Top three right now: Retrace, Pydantic Logfire, Prefactor.
Nobody paid to be on the list and the order is never edited.
Show more
This gives me @marlene_zw vibes on "How to become a billionaire" in the age of AI.
OpenAI will shutoff access to models from Cursor on Nov 12th. Musk's famous contract violations seems to be the reason for this one.
We’re ending our partnership with Cursor following its acquisition by SpaceX. Under our proposal, Cursor’s direct access to our models would end on November 12.
We know that the people most affected by this decision are the developers who rely on OpenAI models in Cursor. We care about their experience in this transition and we’re ready to go above and beyond to support them.
Show more
Congratulations to the @huggingface folks!
May your usage of Monty continue and your OSS work get stronger. Live long and prosper, Hub.
Nvidia has reportedly agreed to buy Hugging Face, the popular open-source AI hub, for $12.9 billion in a move that would let Nvidia both protect its chip empire and jump back into the cloud business.
Show more
Running a voice weather agent with @pydantic ai.
The agent reply ☀️ vs reality 🌨️.
Do evals, folks.
My founder @samuelcolvin sat down with @europetimesnews to chat about how @pydantic went from a side project to 1B+ downloads a month.
Developer experience, open source, and safety for AI agents.
Show more
Today I discovered Old Reddit.
The same as Reddit but old. Cuter illustrations, great for when nostalgia hits.
Want to make $20,000 today? Hack Monty. 🤑
The 3rd and final round of our bounty program for the Monty Python implementation from @pydantic is live now.
Hack the server -> get the secret -> we'll pay you $20,000. Simple.
... unless Monty is actually secure, and you can't hack it, in which case we have a very powerful tool to safely run AI-generated code. 🤞
Show more
@CrusoeAI is now a first-class model provider in Pydantic AI (@pydantic).
Whole open-weight catalog behind one endpoint: GLM, Kimi, Qwen, Nemotron, DeepSeek, Llama, gpt-oss, all one line apart.
Shoutout @DouweM and @lais_bsc
Show more
Hack Monty Round 3 is live.
$20,000 to escape our Rust-based Python sandbox and read a secret off the server it runs on. Now behind a production WebSocket service.
Nobody escaped Round 2. Round 1 fell in under 48 hours.
(probably) the last round before Monty V1.
Show more
Chat GPT's smartest recent move was their partnership with Revolut.
Myself and friends were knees deep into until that pop-up notification.
Airbnb argues the "Biggest Brains" read more than 100 traces before writing a single evaluator. Their rule: an eval written before you've seen the failures mostly measures the author's imagination.
Reading traces means all of them. The database call, the upstream timeout, the browser session that actually broke. Braintrust sees the slice you mail in. Your observability stack sees what happened. The eval score points at the model. The trace is the crime scene.
We implemented Airbnb's whole framework where your system already lives: three layers, judge calibration against human gold sets, production sampling. All of it working code:
Day four: @strawgate remains in charge. Tomorrow: the benchmark.
Show more
“we sandboxed the agent”
meanwhile the agent:
Much like owning a boat the two best days for software developers are:
1) the day it works
2) they day you turn it off for good
We're giving a $5000 referral bonus for anyone who can introduce us to our new GTM Engineer in San Francisco.
The role itself: becoming the face of @zenml_io in the SF AI Engineering community. That means running our own events and meetups, demoing the product to engineering teams, building relationships with builders and buyers, and turning what you hear on the ground into a repeatable GTM motion for us. You need to be technical enough to hold your own in a demo and social enough to fill a room. Definitely not a sit-behind-your-laptop kind of job.
If someone immediately came to mind, please comment your referral here or send me a DM within the next 7 days.
The JD is here:
(yes thats our team cooking)
Show more
