I put up a few PRs to improve prompt cache efficiency actually, to benefit folks using it through API/overages
PROMPT ENGINEERING
-
Strategy game with AI agents as units you prompt
By
–
Very interesting, a strategy game but the units are AI agents that you prompt!
-
Devin AI Achieves Agentic Self-Improvement with One-Shot Implementation
By
–
We have achieved agentic self improvement – i can just copy paste blogposts and tweets into @devinai and it oneshots the complete implementation wasnt actually sure this was gonna work, jaw dropped when it did. this is very out of distribution of the underlying @GoogleDeepMind Gemini Flash Lite model but it Just Worked. Malte Ubl (@cramforce) Mintlify assistant is powered by just-bash with a custom filesystem — https://nitter.net/cramforce/status/2039841201474695333#m
-
Gemma 4 AI/ML API Launch & Giveaway
By
–
AI/ML API now supports Gemma 4 🔥
— 🚨 AI News | TestingCatalog (@testingcatalog) 3 avril 2026
Gemma 4 is one of the strongest open models in terms of speed, cost, and quality, according to early tests. The launch includes an opportunity to win $100 in API tokens for 50 random winners! https://t.co/8YJ34ddx3v pic.twitter.com/wXHG5UtpLxAI/ML API now supports Gemma 4 Gemma 4 is one of the strongest open models in terms of speed, cost, and quality, according to early tests. The launch includes an opportunity to win $100 in API tokens for 50 random winners!
-
Quality Mode: Enhanced World Knowledge and Prompt Understanding
By
–
Deeper World Knowledge Quality mode features dramatically stronger world knowledge and prompt understanding. It accurately interprets complex scenes, realistic physics, object relationships, specific references (brands, locations, culture), and detailed fictional or artistic worlds with precision. (whiteboard by @venturetwins)
-
AGI Systems Should Verify Facts to Avoid Hallucinations
By
–
you could make that argument. or could argue that a system that bills itself as near AGI ought to do a search to check its facts. (does depend a bit on how you define hallucination, i would agree)
-
LangChain Harness Engineering Day 5: Tool Setup and Teardown
By
–
harness eng day 5: toolsets some tools need setup and teardown around the agent loop, like connecting to a tool server or spinning up a sandbox for example, @langchain's ShellToolMiddleware handles init and cleanup, and injects the shell tool into your agent's tool registry!
→ View original post on X — @langchain, 2026-04-03 15:48 UTC
-
The AI Skill That Will Make You Billionaire
By
–
This skill will make you a billionaire. pic.twitter.com/5Y711wgNsh
— Anndy Lian (@anndylian) 3 avril 2026This skill will make you a billionaire.
-
τ³-bench: Interactive Agent Evaluations for Knowledge and Voice
By
–
Really excited for the release of 𝜏³-bench, which brings interactive agent evals ever closer to real-world use cases across two dimensions: 1. 𝜏-knowledge evaluates agents that need to operate over noisy knowledge bases to figure out the correct policies/tools to use while serving a user 2. 𝜏-voice tests voice agents in interactive customer service style settings. If you are developing embedding or voice models for AI agents, 𝜏³ is a great testbed for you to see how your models would perform in a realistic downstream use case. Blog: sierra.ai/blog/bench-advanci… Tweets from @BenShi34 and @keshav_57: nitter.net/benshi34/status/203436… nitter.net/keshav_57/status/20346…
-
Microsoft’s MAI-Image-2 AI Model Generates Creative Images
By
–
Here are few image create using Microsoft's new ai image model – MAI-Image-2 with prompts.