A decade ago, conventional wisdom said custom silicon was a mistake. Today, it is the architectural backbone powering the agentic era. See how ten years of continuous innovation led to our eighth generation TPUs and how we’re building the foundation for what’s next. #8daysofTPU8

Finally @googlecloud Certified DevOps Engineer #4x

Company Brain / Company OS / Agent OS I want this for ME. I'm sick and tired of being the context glue or having to pay the price of over fetching in tokens. But first why me? I run an AI Product Development / R&D Engineering - Hardware and Software in medical. Bitlab.co - We are Mid 7 figure revenue last year. In here our marketing engine is run mostly by AI, and by our CMO (Anatolie ) - Aka the other Glue to the other engine. I operate Health Care AI Agent Company CaesarHealth.com - integrated to large EMRs including AthenaHealth, ModMed, and more partnerships to be announced. We are in multiple acquisition talks with P/E and EMR vendors. I'm a Fractional CTO - Work with multiple startups in Health Care AI - Consulting with also P/E. I can view the stack from CEO, CFO, CIO, CMO, CTO, perspective, and it runs DEEP, but if someone manages to pull this off, it will be MASSIVE VALUE. What started as a massive data mining exercise (create token cache), ended up having to create a new kind of database / datalayer - Fact Layer (because cache gets poisoned by stale) Then I realized once the data is there, you bring the agents. This is Agent Launchpad. All company agents - across the board here, you have visibility into gas (tokens), prompts, skills. If skills is IP, some IP belong in @github (code), but not all of it. Data doesn't belong in github, neither in a relational or just vectore stores. Agents need better data format, You need a data layer, with the MCPs. @claudeai @OpenAI has enough of your data already, @github is the code layer @awscloud @Azure @googlecloud are you runtime. This is why Daiton.ai

Evidence to Action: AI-Powered Insights Across the Drug Development Lifecycle @googlecloud #Pienomial #GoogleCloud #GoogleCloudSpringboard #GoogleCloudPartners google.smh.re/5X_h

I'm building a Company Brain / Company OS / Agent OS - As @naval said, build the company that only you can build. I want this for ME. I'm sick and tired of being the context glue or having to pay the price of that in tokens. But first why me? I run an AI Product Development / R&D Engineering - Hardware and Software in medical. Bitlab.co - We are Mid 7 figure revenue last year. In here our marketing engine is run mostly by AI, and by our CMO (Anatolie ) - Aka the other Glue to the other engine. I also started and run an HealthCare Agent Company CaesarHealth.com - integrated to large EMRs including AthenaHealth, ModMed, and more partnerships to be announced. I'm the CEO here. We are in multiple acquisition talks with P/E and EMR vendors. I'm a Fractional CTO - Work with multiple startups in Health Care AI - Consulting with also P/E. I can view the stack from CEO, CFO, CIO, CMO, CTO, perspective, and it runs DEEP, but if someone manages to pull this off, it will be MASSIVE VALUE. What started as a massive data mining exercise (create token cache), ended up having to create a new kind of database / datalayer - Fact Layer (because cache gets poisoned by stale) Then I realized once the data is there, you bring the agents. This is Agent Launchpad. All company agents - across the board here, you have visibility into gas (tokens), prompts, skills. If skills is IP, some IP belong in @github (code), but not all of it. *more here, will share later. Data is here, Agents are here, You need a data layer, with the MCPs. @claudeai , @OpenAI has enough of your data already, @github is the core layer @awscloud @Azure @googlecloud are you runtime. This is why Daiton.ai

@yikyquintana @ELBETOFERREIRO @Marlins @googlecloud Yo pagué $99 por la subscripcion a Marlins.TV y entonces hoy si no pago la subscripcion a Peacock no puedo ver el juego. 🥴😳

gBunny

this is shaping up to be something: ruggedize.ai — August 26-27 at the farm in Salinas, CA @reservoirfarms @NVIDIARobotics @GoogleDeepMind @googlecloud @JohnDeere @WesternGrowers @OpenRoboticsOrg @awscloud @Agtonomy @carbon_robotics @verdantrobotics @bonsai_robotics @orchard_robots etc.

Kazuma Okamoto got a pitch down and in and sent it UP AND OUT 😳 (MLB x @GoogleCloud)

The Purpose-Driven AI Design Ecosystem hkrampersad.wordpress.com/20… @Oracle $AMD @SEMIAmericas @SEMIconex @AMD @NetApp @NVIDIAAI @nvidia @LisaSu @googlecloud @GoogleAI @OpenAI @deepseek_ai @OpenAINewsroom @sama @ASI_Alliance @IBM @IBMData @drfeifei @elonmusk @DeepSeekR1

Google Cloud が Open Knowledge Format（OKF）を公開——企業知識を Markdown で統一し、AI エージェント直結に。Andrej Karpathy が提唱した「LLM Wiki」を初めて実装するスタンダード。ベンダーロックインなし。 #GoogleCloud #AI #DataManagement 詳細は↓