The way he bolted left many of us to wonder if Alabama was in the works a long time, including how little he thanked UW for hiring him on Montlake and making Tuscaloosa possible. He left in such a hurry it robbed the team of being celebrated for their heroic season

All I can say is this: Over the last seven years I’ve had people call me a “DEI” hire to the FBI. Damn straight: I was (as far as I know) the first Indian American woman in the FBI, who speaks three languages (one that literally no one in the IC existed to even test me for),

Apple says its AI models were trained on Google's custom chips — another proof of the statements I made at @GoogleCloud Next in April during a theCUBE discussion! Following is kif reporting from CNBC 👇🏽 cc John Furrier Dave Vellante Shelly Kramer cnbc.com/2024/07/29/app…

This has been a long, long time coming. We wouldn't even use CodeCommit for re:Invent session content, or for internal projects. It's been a poor option for years. The problem now is lack of trust. AWS has become that which they mocked (Google).

After giving it a lot of thought, we made the decision to discontinue new access to a small number of services, including AWS CodeCommit. While we are no longer onboarding new customers to these services, there are no plans to change the features or experience you get today,

wow, this really is an end of an era. I am sure AWS will take good care of the people using these, and I know Jeff will be a great advocate. but this is a fundamental change that seemed to come from Honeycode several years ago. I say this as an AWS fan, but recognize the change.

"that is the joy that at one point in your life everyone gets to experience". A great friend of mine's baseball team won the state high school championship this year for the first time. he told the kids afterwards "you don't get many dogpiles in your life - remember that one

Gemini 1.5 Pro experimental (0801) is currently ranked #1 on LMSYS for both text and multi-modal!

Yesterday, I ran some tests comparing gemini 1.5 pro exp vs. gpt-4o. Google beat OpenAI, every-single-time. But now, is gemini better than claude? I run the same tests: test #1 → write a viral twitter thread

When you use the 2 million context window in Gemini 1.5 Pro for all the "Dune" books and get answers to complicated questions on themes, tone, etc: Lisan al Gaib! Tune in to learn more about our recent #generativeAI announcements and what they mean for you ↓

Really excited to see us talk publicly about the progress on this work. A few people in our awesome robotics team many years ago came to me and suggested that we should work on this problem (even though it would require buying some pricey higher speed robots) because it would

Today, we are rolling out three experimental models: - A new smaller variant, Gemini 1.5 Flash-8B - A stronger Gemini 1.5 Pro model (better on coding & complex prompts) - A significantly improved Gemini 1.5 Flash model Try them on aistudio.google.com, details in 🧵

Bunch of new models, go give them a try!

Huge gains with 1.5 Flash across the board and 1.5 Pro is much better at math, coding, and complex prompts. Gemini 1.5 Flash is the best most in the world for developers right now ⚡️

Please remain seated for this exclusive behind-the-scenes look at a Google Cloud data center. Learn more about the inner workings of AI Hypercomputer, a supercomputing architecture designed specifically for AI workloads → goo.gle/4dEY9u3

Gemini Flash is now tied with gpt-4o for #2 on the lmsys *vision* leaderboard! combine that with a 1M context length and you can do some seriously cool multimodal work for super cheap ⚡️⚡️⚡️

one of the great things about tech is that what becomes mainstream was often explored by smaller cos- success or not. VERY excited about AirPods hearing aid mode, and wish our team had made it happen at Doppler Labs. HT to David Pierce for this on Doppler labs wired.com/story/inside-t…