Gemma 4 12B Adds Zero‑latency On‑device AI for Desktops
"In tandem with the Gemma 4 12B launch, we are officially introducing powerful on-device developer integrations powered by LiteRT-LM, bringing zero-latency local AI execution natively to standard desktop environments ..." https://t.co/LJXt8b0wfa < dev guide for Gemma 4 12B
Optimizing AI Cold Starts on Cloud Run
A Guide to AI Cold Starts on Cloud Run https://t.co/VmrOVZBN5c < running inference on a serverless stack is pretty amazing. How can you tune performance for faster startup? Good post from @shirmeir86 https://t.co/cqzRqL8cF0
Gemini Managed Agents Blend Hosted Environment with Tool‑Use Loop
"The useful part of Gemini Managed Agents is the combination of a hosted environment and a built-in tool-use loop."
Google Releases Agent Executor, a Potential Industry Standard
Google open-sources Agent Executor to run AI agents in production https://t.co/4BdXW944oJ < We've been intentionally lowkey about this. But @rakyll and team built something that COULD become a standard agent runtime across the industry.
Gemini Adds Versioning and Traffic Splitting for Agents
Worried about overwriting your current agent with a new version and borking something? I am. I like that @googlecloud Gemini Enterprise Agent Platform just added the ability to create agent versions AND do traffic splitting across them. Docs: https://t.co/XXDgds6s5u
Monthly Token Usage Soars to 3.2 Quadrillion
3.2 quadrillion tokens per month. What a shocking increase from just a year ago. #GoogleIO https://t.co/eWMtttdPVd
BigQuery Transforms From SQL to AI‑Driven Graph Platform
Beyond SQL: How BigQuery Evolved into a Complete AI and Graph Platform https://t.co/R897atf0o2 < simplify, simplify. I hope we use products and build architectures that let us strip out complexity. Here's a good example.
Gemini Interactions API Handles Conversation State for You
Terrific guide. You're sleeping on the Gemini Interactions API. Instead of wrestling with stateless APIs where you wrangle all the multi-step conversations/state yourself, it's all managed for you. https://t.co/wvLeEXUM4y
GitLab Reveals New Team Structure and Architectural Vision
GitLab is restructuring their teams and priorities. Their note offers a very clear perspective of the world and the architectural bets they're making. https://t.co/scRo7wcJur
AI Pressure Drives Yum Brands to Overhaul Data Estate
Even if you don't care about all the AI transformation stuff, it does seem as if this executive urgency is forcing companies to finally fix longstanding pain points. Here, Yum Brands is fixing their data estate so that they can get...
AI Floods Code, Outpaces Trust and Triage
"AI increased the supply of code faster than it allowed maintainers to triage contributions. AI didn’t increase the supply of trust yet. That is the real open-source problem: AI is not killing open source. It is killing open contribution as we...
Object Storage Powers AI with Unprecedented Speed
Object Storage is having a moment. This boring, reliable part of a cloud's portfolio matters more now for AI training and inference. @googlecloud Storage Rapid is amazing. 20m queries per second? 15+ TB/s read throughput? Bonkers numbers. https://t.co/d3fWXnuy1k https://t.co/eaemRlmjjD
AI Surge Overloads GitHub While Other Platforms Stay Stable
The Pulse: AI load breaks GitHub – why not other vendors? https://t.co/pExcLarPlT < super solid analysis by @GergelyOrosz here. GitHub is seeing major spikes in usage thanks to AI, but it's not unheard of load. Why was it enough to...
Bad Incentives and Token Maxing Hurt AI Output Quality
"... input quality does matter. And generally, tokenmaxxing degrades input quality." https://t.co/tWrHcqwjBF < more AI is great. But if your incentives are wrong, you won't see the results you want. The "Cobra Effect" is a great story.
Ship Code in Minutes with Gemini CLI Abstraction
Ship code within minutes with the Gemini CLI DevOps Extension https://t.co/gB69hhqVPX < CI/CD has to change, right? Until we figure out what that really looks like, I'll take abstractions that translate intent to infrastructure.