Top stories in AI today: -Open-sourced AI software developer agent
-Apple Vision Pro’s spatial Personas
-Get text-to-speech audio trained on your voice
-Anthropic uncovers 'many-shot jailbreaking'
-6 new AI tools & 4 new AI jobs Read more: http://
therundown.ai/p/open-sourced
-software-developer-agent
…
AGENTS
-
AI breakthroughs: Developer agents, Vision Pro, voice AI, jailbreaking risks
By
–
-
Princeton’s SWE-Agent Turns GPT-4 Into Autonomous Software Engineer
By
–
Researchers from Princeton NLP have developed SWE-agent.
— Rowan Cheung (@rowancheung) 3 avril 2024
It's an open-sourced AI agent that turns GPT-4 into an autonomous software engineer.
The agent achieved accuracy similar to Devin and resolved 12.29% of issues autonomously. https://t.co/LPUtjPaIwaResearchers from Princeton NLP have developed SWE-agent. It's an open-sourced AI agent that turns GPT-4 into an autonomous software engineer. The agent achieved accuracy similar to Devin and resolved 12.29% of issues autonomously.
-
Replit Code Repair Agent Achieves SOTA Debugging Performance
By
–
So much to be excited about at @Replit Developer Day, but Code Repair takes the crown for me.
— Pietro Schirano (@skirano) 2 avril 2024
An agent model specifically designed to debug your code, achieving SOTA performance and even better than GPT-4 and Claude 3 on LeetCode evals.
Great work @amasad @pirroh and the team https://t.co/3JsSsBOEvZSo much to be excited about at @Replit Developer Day, but Code Repair takes the crown for me. An agent model specifically designed to debug your code, achieving SOTA performance and even better than GPT-4 and Claude 3 on LeetCode evals. Great work @amasad @pirroh and the team
-
AI Clone Automates Founder Responses for Investors
By
–
Yup, exactly. When we raised an early funding round, I built an AI version of myself to answer questions automatically. Investors loved it, and it saved me so much time. Amazing that Delphi now exists to do this for everyone.
-
Matt Shumer Launches Personal AI Clone Using Delphi
By
–
Excited to announce that you can now chat with AI Matt. I created an AI clone of myself using @withdelphi
. AI Matt knows quite a bit about training models, writing prompts, and building AI systems. Chat now: https://
delphi.ai/matt-shumer -
Memory Hackathon: San Francisco AI Innovation Event
By
–
Don’t miss our Memory Hackathon in San Francisco this Saturday, in partnership with @AnthropicAI
, @MongoDB
, and
@NewComputer
. Ship new AI memory methods and apps, learn from industry leaders, and win prizes! Register here: -
AI Agent Learns User Writing Style with Self-Improvement
By
–
For our first workshop at The Rundown University, Dr. Cintas created an AI agent that:
— Rowan Cheung (@rowancheung) 2 avril 2024
– Learns a user's writing style
– Writes anything mimicking it
– Improves on its own based on the user's rating of its performance
Then, we dropped the free AI agent code template so anyone… https://t.co/um08F1aMtwFor our first workshop at The Rundown University, Dr. Cintas created an AI agent that: – Learns a user's writing style
– Writes anything mimicking it
– Improves on its own based on the user's rating of its performance Then, we dropped the free AI agent code template so anyone -
Memory Hackathon: Long-term LLM Applications in SF
By
–
Memory Hackathon in SF with LangChain, @newcomputer
, @MongoDB and @AnthropicAI Sign up here: https://
lu.ma/taa6ijxt Long-term memory for LLM applications is one of the most interesting and important – yet also under explored – areas out there. We are THRILLED to be -
Multi-Model AI Architecture: GPT-4 and Claude Haiku Workflow
By
–
The reason I am saying this is that I have been experimenting with this flow: GPT-4 directs and refines instances of Claude Haiku to accomplish a goal. Opus refines and provide the entire output end to end. It's super powerful.