Advait Bopardikar
@AdvaitOnline
Working on Product Strategy for Gemini in Vertex AI. PM @GoogleCloud. All views are my own.
Oh yeah, I forgot that we also made Audio input cheaper to $0.30/1M tokens so its even cheaper to use for use cases like speech transcription and audio time-stamping. developers.googleblog.com/en/gemini-25-f…
Today we are moving Gemini 2.5 Flash-Lite, out most cost-efficient model to GA. You can take advantage of its blazing fast speed, and stronger performance with/without thinking on Vertex AI. See below for more details.
Three new open models in the @googlecloud Vertex AI Model Garden. TimesFM-2.0: Foundation model for time-series forecasting. T5Gemma: Encoder-decoder model with an excellent quality-efficiency trade-off. Gemma 3n: Powerful and efficient open model designed to run locally.
Hot Take: I love planning for complex travel itineraries and it gives me joy to envision what I'm going to do and make the logistics work perfectly. I'm probably not going to use your AI travel agent.
We will be making a version of this Deep Think model available to a set of trusted testers, including mathematicians, before rolling it out to Google AI Ultra subscribers. Stay tuned 😉♊
Drastic progress on maths with Gemini 2.5! As a math undergrad, I am impressed 🤯 🥈 -> 🥇 ✅ Formal -> Informal ✅ Specialized model -> General model ✅ Available soon ✅ Huge thanks to IMO and congrats to all participants! Blog: deepmind.google/discover/blog/…
🚀 More innovation to come and way more adoption. things accelerated again once we GA'd 2.5 in June so look forward to the next edition of this chart.
Model demand change 2024 to 2025: Google (+49pts), DeepSeek (+53pts) and xAI (+31pts) have achieved massive gains in demand share over the past year @Google has transitioned from being an AI laggard to an AI leader with a ~2.5x increase in proportion of respondents using or…
4/5 Search now has the agentic capability to call local businesses using AI to check on prices & availability, saving you the hassle of tracking down information yourself. This is rolling out in the U.S., with increased access for AI Pro & AI Ultra subscribers. Makes it much…
I like growth. Only way is up and we have more to come.
Gemini's period-over-period growth has significantly accelerated.
📣 Vertex AI Memory Bank is now supported in the Agent Development Kit! The Agent Development Kit (ADK) now integrates with Vertex AI Memory Bank on Agent Engine, a managed memory service to help you create memory-enabled agents with just 4 steps: 1️⃣ Create an Agent Engine 2️⃣…
Want to make it easy to add Memory to your AI Agents. With the new Vertex AI Memory Bank you now can!
Providing your Agents with long-term memory is often a hassle. Thats why we are introducing Vertex AI Memory Bank. Memory Bank gives your AI agents long-term memory. ✅ Managed API (no more DIY) 📈 Scales beyond context windows 🧩 Integrates with ADK, LangGraph and CrewAI…
Gemini Deep Research has gotten very good since it was upgraded to 2.5 Pro. Claude and ChatGPT tend to go a good job acting like analysts and building an argument, but Gemini is probably the best right now at synthesizing a coherent overview of a complex topic.
Now everyone can access the Veo3 API in Vertex and embed it's functionality into their applications. Go Build, Create and Direct!
Veo 3 is now available in public preview on #VertexAI 🎉 Check out what Google Cloud customers and partners are creating ↓ 🧵
"Google ships their org chart, LOL" was some feedback after seeing the auth options in the Gemini CLI. We've already simplified this, and it reflects dev scenarios: (a) free and easy, (b) higher limits, (c) enterprise-needed control. github.com/google-gemini/…
LLMs are cool but there is more to AI than just language. AlphaGenome: AI for better understanding the genome - Google DeepMind share.google/0SMoubDALD7KVq…
🎉 DeepMind just dropped AlphaGenome: AI that predicts gene regulation from DNA sequence. This is a monumental step towards cracking the genome's 'software code,' unlocking precision medicine & drug discovery 🧬. And it's available via API!