Phil Jund
@philippjund.bsky.social
Advancing LLMs. I love running, cycling, calisthenics. Passionate about sustainability and climate. Previously: fore ai, On-device ML @ Google; Uni Freiburg
Reposted by Phil Jund
DeepSeek is a side project 🔥
January 24, 2025 at 8:01 AM
DeepSeek is a side project 🔥
Reposted by Phil Jund
I have spent a lot of time with AI agents (including Devin and Claude Computer Use) and they really do remain too fragile & not "smart" enough to be reliable for complicated tasks.
So you have two options: (1) wait for better models or (2) focus on narrower use cases (like Gemini Deep Research).
So you have two options: (1) wait for better models or (2) focus on narrower use cases (like Gemini Deep Research).
January 18, 2025 at 5:54 AM
I have spent a lot of time with AI agents (including Devin and Claude Computer Use) and they really do remain too fragile & not "smart" enough to be reliable for complicated tasks.
So you have two options: (1) wait for better models or (2) focus on narrower use cases (like Gemini Deep Research).
So you have two options: (1) wait for better models or (2) focus on narrower use cases (like Gemini Deep Research).
Reposted by Phil Jund
I believe the European standard unit for fuel efficiency is litres per 100km, which is a unit of area. 🤯
January 15, 2025 at 2:55 PM
I believe the European standard unit for fuel efficiency is litres per 100km, which is a unit of area. 🤯
Reposted by Phil Jund
I'm excited about TabPFN for energy forecasting! It's a new ML model for tabular data which is (almost) a drop-in replacement for XGBoost. It's transformer-based, and pre-trained on millions of synthetic datasets.
Code: github.com/PriorLabs/Ta...
Paper: www.nature.com/articles/s41...
1/4
Code: github.com/PriorLabs/Ta...
Paper: www.nature.com/articles/s41...
1/4
GitHub - PriorLabs/TabPFN: ⚡ TabPFN: Foundation Model for Tabular Data ⚡
⚡ TabPFN: Foundation Model for Tabular Data ⚡. Contribute to PriorLabs/TabPFN development by creating an account on GitHub.
github.com
January 13, 2025 at 12:08 PM
I'm excited about TabPFN for energy forecasting! It's a new ML model for tabular data which is (almost) a drop-in replacement for XGBoost. It's transformer-based, and pre-trained on millions of synthetic datasets.
Code: github.com/PriorLabs/Ta...
Paper: www.nature.com/articles/s41...
1/4
Code: github.com/PriorLabs/Ta...
Paper: www.nature.com/articles/s41...
1/4