Last chance to buy early-bird tickets for Interrupt — and we've got more speakers joining the lineup! Sourabh (
@sshirhatti
) and Matas (
@MatasRastenis
) from @uber will be sharing how they're building reusable agent components with LangGraph. Expect real production examples
CODE
-
Uber engineers share LangGraph reusable agent components
By
–
-
Frontier LLMs Earn $1M Real-World Freelance Software Engineering
By
–
SWE-Lancer: Can Frontier LLMs Earn $1 Million from Real-World Freelance Software Engineering? Miserendino et al.: https://
arxiv.org/abs/2502.12115 #Artificialintelligence #Deeplearning #MachineLearning -
Open Deep Research Agent with Configurable LLM Planning
By
–
Open Deep Research Deep research is one of the most popular agent use-cases. Here is an open deep researcher w/ ability to configure report structure, planner/writer LLMs, search APIs, search depth, etc. https://
youtu.be/2mSNIX-l_Zc Code: https://
github.com/langchain-ai/o
pen_deep_research
… -
Auto-manage context optimization with caching for efficient token usage
By
–
We’ve also adjusted how auto-manage context works. When caching is active, more of your chat history is included while keeping total points lower. This ensure you get more context for less. (3/5)
-
Chat History Caching Reduces API Usage and Costs
By
–
We're launching chat history caching to help reduce point usage. If you send messages within five minutes of each other, the model can reuse parts of your chat history instead of starting from scratch. This reduces processing and saves you points. (1/5)
-
/repeat feature enables multiple bot responses per prompt
By
–
New feature: /repeat. Get multiple responses from any bot by starting your message with /repeat, followed by a number and your prompt. (1/2) pic.twitter.com/Nirq62eqpl
— Poe (@poe_platform) 20 février 2025New feature: /repeat. Get multiple responses from any bot by starting your message with /repeat, followed by a number and your prompt. (1/2)
-
GPT vs T5 BART: Decoder vs Encoder-Decoder Models
By
–
GPT (Only Decoder) → Best for text generation T5, BART (Full Encoder-Decoder) → Best for translation & summarization
-
Making Tinygrad a PyTorch Backend for Frontend Stickiness
By
–
i think you should make tinygrad a pytorch backend, frontend support and stickiness is a large part of why pytorch is successful and that's not easy to replicate
-
PyTorch Derived from Torch: Index Return Differences
By
–
i think you keep forgetting pytorch derived from torch, not numpy.
we fixed as much differences as we can, but torch always returned indices in torch -
Ultrascale Playbook: Hugging Face Spaces Resource Guide
By
–
link: link: https://
huggingface.co/spaces/nanotro
n/ultrascale-playbook
…