Petr Brzek
@petrbrzek.bsky.social
We're back with Prague Gen AI Meetup in 2 days.
Two solid talks:
• Peter Bokor (Callstack AI) on building coding agents
• Hynek Kydlíček (Hugging Face) on training small LLMs
Doors at 18:30, talks at 18:45. Beer and networking throughout.
No hype, just tech talk.
Two solid talks:
• Peter Bokor (Callstack AI) on building coding agents
• Hynek Kydlíček (Hugging Face) on training small LLMs
Doors at 18:30, talks at 18:45. Beer and networking throughout.
No hype, just tech talk.
January 28, 2025 at 9:05 PM
We're back with Prague Gen AI Meetup in 2 days.
Two solid talks:
• Peter Bokor (Callstack AI) on building coding agents
• Hynek Kydlíček (Hugging Face) on training small LLMs
Doors at 18:30, talks at 18:45. Beer and networking throughout.
No hype, just tech talk.
Two solid talks:
• Peter Bokor (Callstack AI) on building coding agents
• Hynek Kydlíček (Hugging Face) on training small LLMs
Doors at 18:30, talks at 18:45. Beer and networking throughout.
No hype, just tech talk.
I personally love Deepseek R1 so far. Great for coding and harder reasoning tasks. Now it's easy to fiddle with from Langtail. 🫡
January 24, 2025 at 9:38 AM
I personally love Deepseek R1 so far. Great for coding and harder reasoning tasks. Now it's easy to fiddle with from Langtail. 🫡
What’s the biggest thing you’re missing to do great LLM evals?
January 1, 2025 at 6:17 PM
What’s the biggest thing you’re missing to do great LLM evals?
🚀 Just added DeepSeek-V3 to Langtail!
This beast was trained on ~15T tokens and really shines at coding and following instructions. Impressive how it keeps up with (and often beats) the big closed-source models.
Try it out in Langtail 💅
This beast was trained on ~15T tokens and really shines at coding and following instructions. Impressive how it keeps up with (and often beats) the big closed-source models.
Try it out in Langtail 💅
December 28, 2024 at 11:17 AM
🚀 Just added DeepSeek-V3 to Langtail!
This beast was trained on ~15T tokens and really shines at coding and following instructions. Impressive how it keeps up with (and often beats) the big closed-source models.
Try it out in Langtail 💅
This beast was trained on ~15T tokens and really shines at coding and following instructions. Impressive how it keeps up with (and often beats) the big closed-source models.
Try it out in Langtail 💅
Gotta say, when building AI apps, my favorite combo is Vercel's AI package paired with Langtail.
December 26, 2024 at 4:09 PM
Gotta say, when building AI apps, my favorite combo is Vercel's AI package paired with Langtail.
Was planning to release this for last-minute gift shoppers like myself, but I won't even manage to release it in time - it's already time to celebrate Christmas. Maybe some other time. 🎄
December 24, 2024 at 2:04 PM
Was planning to release this for last-minute gift shoppers like myself, but I won't even manage to release it in time - it's already time to celebrate Christmas. Maybe some other time. 🎄
🚀 OpenAI's new o1 models are now available in Langtail!
December 23, 2024 at 4:00 PM
🚀 OpenAI's new o1 models are now available in Langtail!
🔥 Just added Gemini 2.0 Flash (Experimental) to Langtail!
First model we've tested that actually nails object bounds detection in images. No more guessing - this thing is precise.
Been testing it extensively and honestly, we're impressed.
First model we've tested that actually nails object bounds detection in images. No more guessing - this thing is precise.
Been testing it extensively and honestly, we're impressed.
December 16, 2024 at 4:00 PM
🔥 Just added Gemini 2.0 Flash (Experimental) to Langtail!
First model we've tested that actually nails object bounds detection in images. No more guessing - this thing is precise.
Been testing it extensively and honestly, we're impressed.
First model we've tested that actually nails object bounds detection in images. No more guessing - this thing is precise.
Been testing it extensively and honestly, we're impressed.
Just launched a simple tool to compare different LLMs.
You can filter by context size, cost per token, vision features, and function calls. Pretty useful when you're trying to pick the right model for your project.
langtail.com/llm-price-c...
You can filter by context size, cost per token, vision features, and function calls. Pretty useful when you're trying to pick the right model for your project.
langtail.com/llm-price-c...
December 13, 2024 at 4:38 PM
Just launched a simple tool to compare different LLMs.
You can filter by context size, cost per token, vision features, and function calls. Pretty useful when you're trying to pick the right model for your project.
langtail.com/llm-price-c...
You can filter by context size, cost per token, vision features, and function calls. Pretty useful when you're trying to pick the right model for your project.
langtail.com/llm-price-c...
Made Langtail a bit nicer today - faster prompt/test switching and a cleaner sidebar. The kind of small improvements that make me happy. 🎯
December 11, 2024 at 7:02 PM
Made Langtail a bit nicer today - faster prompt/test switching and a cleaner sidebar. The kind of small improvements that make me happy. 🎯
We've just added support for the new Gemini-Exp-1206 model, currently ranked first on an LLM leaderboard. Try it out in Langtail!
December 10, 2024 at 3:53 PM
We've just added support for the new Gemini-Exp-1206 model, currently ranked first on an LLM leaderboard. Try it out in Langtail!
Your prompts probably suck. (Hard truth 😅)
Test them against PromptNinja, our free tool that fights your prompts with unexpected inputs. See how many battles you win! 🥷
Link in the comments.
Test them against PromptNinja, our free tool that fights your prompts with unexpected inputs. See how many battles you win! 🥷
Link in the comments.
December 5, 2024 at 9:20 PM
Your prompts probably suck. (Hard truth 😅)
Test them against PromptNinja, our free tool that fights your prompts with unexpected inputs. See how many battles you win! 🥷
Link in the comments.
Test them against PromptNinja, our free tool that fights your prompts with unexpected inputs. See how many battles you win! 🥷
Link in the comments.
The Langtail team is at YC Startup School in London. If you want to chat about evaluating LLMs, come and say hi.
November 23, 2024 at 11:14 AM
The Langtail team is at YC Startup School in London. If you want to chat about evaluating LLMs, come and say hi.