huge day for the company that invented web search
We’re partnering with Google Cloud to power Gemini agents with data from the web and beyond.
Through Alexandria, agents get the live web plus official data providers and specialized indexes in one place.
firecrawl.dev/integrations/g…
ZDR used to mean something pretty specific but now you have to decipher a maze of documentation and fine print around request logs, abuse monitoring and everything else just to figure out what the guarantee actually covers
every Bedrock docs page I read somehow left me with more questions
the best harness benchmark is how much of the harness you can delete
one of the amusing things about AI is that infra and security spent a decade getting everyone off long-lived static API keys and onto short-lived tokens
then OPENAI_API_KEY happened and we collectively forgot all of it
I get that managed identity and OAuth are harder but we really did speedrun back to secrets in .env
years of cryptography lore resolved by “lol yeah they probably just called rand()”
CLAUDE FOUND THE SEEDS
anthropic.com/research/ecdsa…
You’re right. That was my mistake.
Anyway, I’ve opened 14 files, spawned 6 subagents, written a 2,300-token implementation plan and asked one agent to challenge the premise of the one-line change.
the true Turing test is surviving the 401 at rush hour
We’re bringing the magic of Waymo to Toronto! 🇨🇦
In the coming days, you’ll begin seeing Waymo’s all-electric vehicles driving through Toronto’s streets as we lay the groundwork for our ride-hailing service.
Our goal is to begin serving riders in 2028.
Sign up for future updates: waymo.com/waymo-in-canada
zuck in prod tweaking css he just like me frfr
Muse Spark Open -- weights releasing today -- runs great on new Mac Studios thanks to work with @johnternus and Apple.
wait till Devin Oncall discovers labor law and enters its misalignment arc
Scout’s hardest eval: explain the Datadog bill
Meet Scout 1.0, our first AI model built for investigating production incidents.
Scout matches frontier-model investigation quality at a fraction of the cost, which makes it practical to investigate more issues, earlier.
We trained Scout on realistic simulations built from real-world incidents and production telemetry, using reinforcement learning to reward accurate root causes backed by evidence.
Learn more and sign up for the preview: bit.ly/3TJknGO
Paxos crawled so “Claude, make no mistakes” could sprint
Do you remember when @bcherny said Claude couldn't build distributed systems?
We have a fix for that. 🫡
antithesis.com/blog/2026/tim…
every date is now a red-team exercise
OPENAI to EMPLOYEES: If someone is trying to seduce you, first consider — could they be a spy???
OpenAI is warning its employees about honeypot spies and strangers with children in playgrounds ... a very fun one with @georgia_wells and @MeghanBobrowsky:
wsj.com/tech/ai/ai-companies…
superintelligence is nice and all but can it pressure test the big rocks without boiling the ocean and circle back once there’s alignment
a lot of the agent architecture from a year ago was compensating for models being dumb.
the models improved faster than people’s willingness to remove the scaffolding, especially once they became emotionally attached to the abstractions they built
