Katie Watson

Comms, Google Cloud. Enterprise hardware & software junkie, lover of pretty stuff and places.

[Demis Hassabis @demishassabis]

For 20+ years @ShaneLegg and I've discussed AGI’s potential impact on the economy, science & society. With the DeepMind Institute, we're expanding interdisciplinary research on key questions for the AI era. We hope it spurs the discussions needed to get the next steps right: https://t.co/Bqf10G3Xep

[Sundar Pichai @sundarpichai]

Introducing Gemini 3 ✨

It’s the best model in the world for multimodal understanding, and our most powerful agentic + vibe coding model yet. Gemini 3 can bring any idea to life, quickly grasping context and intent so you can get what you need with less prompting. 

Find Gemini 3 Pro rolling out today in the @Geminiapp and AI Mode in Search. For developers, build with it now in @GoogleAIStudio and Vertex AI. 

Excited for you to try it!

[Sundar Pichai @sundarpichai]

Our 7th gen TPU Ironwood is coming to GA! 

It’s our most powerful TPU yet: 10X peak performance improvement vs. TPU v5p, and more than 4X better performance per chip for both training + inference workloads vs. TPU v6e (Trillium). We use TPUs to train + serve our own frontier models, including Gemini, and we’re excited to make the latest generation available to @googlecloud customers.

[Rohan Paul @rohanpaul_ai]

Google Cloud's rise is an amazing story.

⚙️ Google Cloud flipped from laggard to growth driver on AI demand, posting $15B Q3 revenue at 34% growth in Oct-25, and now rivaling YouTube as Alphabet’s #2 cash source.

The engine is a companywide shift toward enterprise discipline plus years of investment in datacenters, custom tensor processing unit chips, and faster networking.

Market share rose from 7% in 2018 to 13% in 2025, while Microsoft is near 20% and Amazon near 30%, so it is still smaller but closing.

Profitability arrived in 2023 after losses from 2018 to 2022, with sales reorganized by industry and targets reset to revenue rather than bookings.

In 2022 Thomas Kurian moved TPU commercialization into Cloud and opened access beyond internal teams, letting partners and even rivals rent Google silicon directly through the platform.

That helped sign 9 of 10 leading AI labs and pushed Anthropic to plan up to 1M TPUs, while others like Apple and Safe Superintelligence also adopted the chips.

Google launched an enterprise version of Gemini in Oct-25 and signaled willingness to host other model families, which gives large customers more choice.

The bill is heavy with $91B to $93B capital spending planned for 2025 after a step up from $85B, and leadership expects an even larger build in 2026.

For Google, YouTube still delivers scale with 1B hours watched per day, yet Cloud now carries more weight in Alphabet’s planning and resource debates.

---

reuters .com/business/ai-turned-google-cloud-also-ran-into-alphabets-growth-driver-2025-10-31/

[Katie Watson @kwatson49]

The future of high-performance cluster management is here.

@GoogleCloud is simplifying access to powerful infrastructure with our new managed Slurm service, offering a seamless experience for massive AI workloads.
https://t.co/VmBEsrPE2e

[ryanjsalva @ryanjsalva]

I love the folks at @zeddotdev so much. ❤️ If you haven't tried it, Zed is possibly the world's best code editor for speed and collaboration. We approached them a few weeks ago looking for an opportunity to collaborate, and together we dreamed up this integration that treats the Zed as a remote agent for Gemini CLI. Among other things, you can:

- Take your Gemini CLI (Code Assist) quota with you to Zed
- Follow Gemini CLI in Zed as it makes edits
- Ask Gemini CLI questions about logs and errors in Zed

This won't be the end of our collaboration. It's one heck of a beginning. Thanks, Frederic Lardinois at @thenewstack for chatting with us and sharing with the world.

https://t.co/gndelP0IiZ

[Google AI @GoogleAI]

Gemini 2.5 Flash-Lite is now stable and generally available for developers and enterprise customers! ⚡

When designing a Gemini model, we think a lot about the tradeoffs between quality, cost, and latency. Previously with 2.0 Flash-Lite we optimized for cost-efficiency over latency. As we built our next iteration, we also wanted to push the boundaries on latency to see how fast we could get the model to think and respond.

Resulting in 2.5 Flash-Lite, our fastest, most cost-efficient 2.5 model yet, with lower latency than both 2.0 Flash-Lite and 2.0 Flash on a broad sample of prompts.

Try it out in https://t.co/7pwDnyTTNL and @GoogleCloud Vertex AI.

[Katie Watson @kwatson49]

@nytimes on Google's Big Sleep AI agent news, what the adversary is really doing with AI per @JumpforJoyce, https://t.co/6IcEDB3xg8

[Sundar Pichai @sundarpichai]

New from our security teams: Our AI agent Big Sleep helped us detect and foil an imminent exploit. We believe this is a first for an AI agent - definitely not the last - giving cybersecurity defenders new tools to stop threats before they’re widespread.

[Google Cloud Tech @GoogleCloudTech]

Gemini CLI ❤️ your ⭐⭐⭐

A huge thank you to everyone around the world contributing to this new open source project.

If you haven’t already, come build with us → https://t.co/ED2bDraA7d

[Julian Goldie SEO @JulianGoldieSEO]

NEW Gemini CLI Update 🤯 https://t.co/IG8ucN92fA

[Rohan Paul @rohanpaul_ai]

Insane speed in research release by Google.

The latest one AlphaGenome, will help understand how our genes work.

AlphaGenome is a deep-learning “sequence-to-function” model that combines convolutional layers and a transformer tower to process 1 Mb of DNA and predict thousands of genomic tracks.

🧬 The Core Idea

AlphaGenome feeds 1 million letters of raw DNA into a U-Net style encoder-transformer-decoder.

Early layers spot short useful patterns.

Middle transformer layers let far-apart sections share information.

Final layers give precise predictions for every letter or small block.

This design keeps single-letter detail while still seeing distant control regions.

[Katie Watson @kwatson49]

@ntaylormullen Congratulations, Taylor!!! Huge

[Richard Seroter @rseroter]

We'll likely never tell this story fully, but this effort was a startup within Google, by engineers for engineers.

It was a hungry, focused, brilliant team of folks like @ntaylormullen, @allen_hutchison, @thechrisperry, @rakyll, @ryanjsalva and other stars. Super fun.

[Logan Kilpatrick @OfficialLoganK]

Say hello to the @geminicli, a local CLI to help you build and maintain software with 1,000 free Gemini 2.5 Pro requests per day : )

[The Linux Foundation @linuxfoundation]

Breaking news 📢 The Linux Foundation launched the Agent2Agent Protocol project at #OSSummit!

Created by @Google, A2A enables secure, open communication between AI agents across vendors, platforms and frameworks to improve modularity, mitigate vendor lock-in and accelerate innovation.

More on the growing collaborative ecosystem here: https://t.co/IVwqo1PMxW

[François Chollet @fchollet]

Gemini 2.5 Pro has been my default for a few weeks. Best model I've used so far

[Google Cloud @googlecloud]

We released the Agent2Agent (A2A) protocol as #OpenSource and are excited to see today's news from @Microsoft to support A2A! 🎉

Join the A2A protocol party and help us define the future of agent interoperability → https://t.co/M6ntIJ9Mf7

[Katie Watson @kwatson49]

Microsoft adopts #Google’s standard for linking up AI agents https://t.co/AT4fbcrLNf via @techcrunch

[Eli Lifland @eli_lifland]

A year before ChatGPT Daniel wrote "What 2026 Looks Like" which foretold the rise of chatbots, chain-of-thought, inference scaling, and more. But it stopped before AGI.

Now I've worked with him to write a sequel. Read AI 2027 to see how AI takeover could actually happen.