We make AI models Dolphin and Samantha
BTC 3ENBV6zdwyqieAXzZP2i3EjeZtVwEmAuo4
https://t.co/3ri2GbWU13
https://t.co/zH0F3pSLuq @dphnAIerichartford.com Charlotte, NCJoined October 2014
Just had an aha moment w/ GEPA @DSPyOSS
- Gemini 2.5 Flash-Lite (GPT 4.1 reflection LM)
- 3 signatures in one compound module
- 12 training examples
- 10 test examples
- 32 minutes of optimizer runtime
- $0.90 total cost
Results:
- Baseline: 68.2%
- GEPA-Optimized: 95.3%
🤯
Recent models tested on longform writing:
Sonoma-sky-alpha: cloaked model appears to be grok.
Qwen3-max: has the same long context degradation issue as qwen3-235b: converges on super short 2-3 word paragraphs.
Kimi-k2-0905: slightly worse than k2 but ~within margin of error
VS Code adds support for custom OAI-compatible endpoints
This a big win for local AI as it allows us to use any local model provider without vendor lock-in. Big thanks to the VS Code devs and especially @IsidorN for listening to the community feedback and adding this option!
But, how does the original Qwen3-14B do on AIME24, AIME25, and HMMT25? (they didn't put it in the chart) It would be cool to see how much the rStar2-Agent training improved the original model.
But, how does the original Qwen3-14B do on AIME24, AIME25, and HMMT25? (they didn't put it in the chart) It would be cool to see how much the rStar2-Agent training improved the original model.
If you can build a tinybox for a lot less, you should start a competing business!
How do you square the "but I added up the parts" with the fact that tinybox is the cheapest off the shelf option?
If you can build a tinybox for a lot less, you should start a competing business!
How do you square the "but I added up the parts" with the fact that tinybox is the cheapest off the shelf option?
Grok Code Fast from @xai scored 90% on Roo Code evals — top-tier performance at half the cost of its peers. ⚡️
Free to try in Roo Code Cloud until Sept 10. See why speed + savings make @grok a strong new addition: roocode.com/evals
50 Followers 467 FollowingA great t-shirt description should be clear, concise, and highlight the shirt's best features and benefits.
Tap the link in our bio...
148 Followers 1K FollowingFounder & CEO of Flowerpilot. Hooked on product engineering, renewable energy, the future of Europe, and the Chinese economy.
2K Followers 5K FollowingWe help companies apply AI and data science to real-world problems in biology, healthcare, and tech. From genomics to automation, we ship.
268 Followers 685 Followingcollaboration [email protected]
Organic nonllm AI in validation phase. Multi modal tokenizer 550k tokens per second and only 34 mb of ram
205K Followers 5K FollowingVC at @MenloVentures. Formerly founding team @glean, @Google Search. @Cornell CS. Tweets about tech, immigration, India, fitness and search.
4K Followers 7 FollowingThe Decentralized, Distributed Serverless AI Compute Platform, by @rayon_labs.
Powering https://t.co/zkoLw8OPwb, https://t.co/3m1GjXs2Tg and your next AI App.
17K Followers 20 FollowingA high-throughput and memory-efficient inference and serving engine for LLMs. Join https://t.co/lxJ0SfX5pJ to discuss together with the community!
8K Followers 3K Following🤖 Business simulators at https://t.co/QtPqoTk2G0
📕 O'Reilly author on Prompt Engineering
🎓 500k students have taken my courses
📈 Built a 50 person growth agency
17K Followers 21 FollowingAn AI research and product company 🫠. We are a team of scientists and engineers building state-of-the-art multimodal models 😻
6K Followers 213 FollowingFrench Robotics company
Doing open source robots made by open minded humans to explore real-world applications
Join our community : https://t.co/8j2MrPPHzG
41K Followers 1K Followinghttps://t.co/6gcBkfdPnt | Book: 101 Things All Young Adults Should Know | Right Wing News Founder | Raised 600k in a GoFundMe for Brett Kavanaugh |
11K Followers 34 Following(an alien👾) GLSL artist/ PhD @tudelft/ Simulation Theory/ Grand Award @ADAAman2020/ JACK James Award @ArtOlympia/ My code shared @x is MIT Licensed (🦠西辻󠄀 陽平)