Mikhail Sirotenko
@msirotenko.bsky.social
TLM@ Google DeepMind working on multimodal AI. Previously stealth startup < Amazon < Samsung < another startup < PhD.
Reposted by Mikhail Sirotenko
🥁Introducing Gemini 2.5, our most intelligent model with impressive capabilities in advanced reasoning and coding.
Now integrating thinking capabilities, 2.5 Pro Experimental is our most performant Gemini model yet. It’s #1 on the LM Arena leaderboard. 🥇
Now integrating thinking capabilities, 2.5 Pro Experimental is our most performant Gemini model yet. It’s #1 on the LM Arena leaderboard. 🥇
March 25, 2025 at 5:25 PM
🥁Introducing Gemini 2.5, our most intelligent model with impressive capabilities in advanced reasoning and coding.
Now integrating thinking capabilities, 2.5 Pro Experimental is our most performant Gemini model yet. It’s #1 on the LM Arena leaderboard. 🥇
Now integrating thinking capabilities, 2.5 Pro Experimental is our most performant Gemini model yet. It’s #1 on the LM Arena leaderboard. 🥇
Reposted by Mikhail Sirotenko
There are going to be Stand Up for Science events this Friday in DC and around the country. Gotta do what we can to be heard.
On Friday, we Stand Up for Science in locations Nationwide (& internationally!) to support science as a public good that benefits everyone.
Join us! Visit www.standupforscience2025.org for more information & other ways to get involved ☀️
Join us! Visit www.standupforscience2025.org for more information & other ways to get involved ☀️
March 4, 2025 at 5:31 PM
There are going to be Stand Up for Science events this Friday in DC and around the country. Gotta do what we can to be heard.
Reposted by Mikhail Sirotenko
We’ve been thrilled by the positive reception to Gemini 2.0 Flash Thinking we discussed in December.
Today we’re sharing an experimental update w/improved performance on math, science, and multimodal reasoning benchmarks 📈:
• AIME: 73.3%
• GPQA: 74.2%
• MMMU: 75.4%
Today we’re sharing an experimental update w/improved performance on math, science, and multimodal reasoning benchmarks 📈:
• AIME: 73.3%
• GPQA: 74.2%
• MMMU: 75.4%
January 22, 2025 at 12:31 AM
We’ve been thrilled by the positive reception to Gemini 2.0 Flash Thinking we discussed in December.
Today we’re sharing an experimental update w/improved performance on math, science, and multimodal reasoning benchmarks 📈:
• AIME: 73.3%
• GPQA: 74.2%
• MMMU: 75.4%
Today we’re sharing an experimental update w/improved performance on math, science, and multimodal reasoning benchmarks 📈:
• AIME: 73.3%
• GPQA: 74.2%
• MMMU: 75.4%
The blogpost is out about our recent work on training masked autoencoders on long(-er) videos. The paper was accepted to NeurIPS`24.
More at: goo.gle/4fW5aIc
More at: goo.gle/4fW5aIc
Extending video masked autoencoders to 128 frames
goo.gle
December 4, 2024 at 11:03 PM
The blogpost is out about our recent work on training masked autoencoders on long(-er) videos. The paper was accepted to NeurIPS`24.
More at: goo.gle/4fW5aIc
More at: goo.gle/4fW5aIc
Reposted by Mikhail Sirotenko
So how long do we give it before the bots and spam arrive here, and how well do we think @bsky.app is prepared?
1. L I N K I N B I O.
2. Crypto spam
3. If you’re not doing these 10 GenAI things, you’re falling behind
4. Wow. This changes everything.
My guesses: 2 months, and not well.
1. L I N K I N B I O.
2. Crypto spam
3. If you’re not doing these 10 GenAI things, you’re falling behind
4. Wow. This changes everything.
My guesses: 2 months, and not well.
November 24, 2024 at 8:24 PM
So how long do we give it before the bots and spam arrive here, and how well do we think @bsky.app is prepared?
1. L I N K I N B I O.
2. Crypto spam
3. If you’re not doing these 10 GenAI things, you’re falling behind
4. Wow. This changes everything.
My guesses: 2 months, and not well.
1. L I N K I N B I O.
2. Crypto spam
3. If you’re not doing these 10 GenAI things, you’re falling behind
4. Wow. This changes everything.
My guesses: 2 months, and not well.