Follow AI Builders — Not Influencers

RT ollama Ollama v0.33 is here! You can now easily configure Claude Desktop to seamlessly work with Ollama as a third-party gateway provider. One toggle. Cloud & local models just work👇

Deepagents is becoming a multiplayer harness Auth, memory, etc If you’re building an agent that you want to expose to multiple users in same thread and are thinking about these issues - reach out! Would love to chat

For the first time ever, I feel like I can trust AI to be my project manager... for basically everything. Now that I have agents working in the background 24/7, with context + plugins connected to my entire company... I’m able to relax and focus on the task at hand. I work with six different agencies that operate across different channels, Slack, email, Notion, text, WhatsApp, ClickUp, and more. It used to absolutely drain me to keep track of everything... But now, I have AI keeping track of it all. And I’m slowly giving it the ability to respond to specific types of emails and Slack messages without asking for permission. Every night, it adds all of that context to an internal company brain, which is basically one giant Notion database... the docs for my entire company. I can plug any agent into this system. Then, every month, I have a mega-prompt where Fable 5 goes through everything, trims the fat, and gets rid of duplicates before I manually review it. It’s like my monthly company review. It focuses on SOPs and company goals, and then I manually write my goal for the next month... which helps guide responses from every agent. In summary: 1. Give your agents context on your entire company 2. Constantly remind your agents of your goals 3. Slowly give them more freedom to take action 4. Do a monthly review with the most powerful model available to realign everything

A background process that continually ingests context from every single connector (or app), performs multi-hop reasoning in a perpetual inference loop, and runs on your hardware. That is the future.

I'm actually curious if it's that scary to be in the #1 camp * I still think humans will use software instead only doing work through agent apps. So humans will continue to use SaaS through the UI and agents will use SaaS through the APIs * as other folks have mentioned in the comments, users likely won't do general agentic work through this software, but they'll make use of specialized agents for specific tasks * software can charge per seat and also with a credit allocation on API calls * It takes real work to maintain and improve software that's good enough to run your production workflows through it. If you DIY your own system of record, you'll be eating human or token costs trying to maintain it for your evolving needs, which is basically the same thing as paying a subscription