@kunchenguid Every harness will abstract this away from the user anyway. The Jev primitive will resolve this! Including when to know if the user is “done” and keep the cache alive for the user
@dillon_mulroy Yes, it is filling this missing piece that many of us stuck in smaller and fast LLMs instead but it was never quite right. Half of all fine tuned models just became obsolete
I must be among an extremely small group of people (n=1?) that have both 1) trained a frontier LLM and 2) designed and synthesized custom viruses in a lab with my own two hands.
And I think that the takes on AI killing us all by creating dangerous viruses is total bogus.
@RhysSullivan We’re basically doing this, but in a managed way with configuration. But will keep trending towards just giving all clients access to everything on the front end.
Some AI agent terms and how I think of it:
- Skill: can be anything that is progressively disclosed. docs, instructions, memory, etc.
- Playbook: Is automatically ran based on some state or intent, not optional like a skill. Typically will be step by step guide for agent based on some workflow. Can explicitly list out skills that should be used.
- Command: Can be a Skill or Playbook that is explicitly invoked by a user
@bscholl True, they could have just said: here is a phone that expands so you have more space to doom scroll, watch TikTok, and chat with friends while sitting on your coach. Sold!
@RhysSullivan@thdxr This is what I built into our harness. Can use code mode through a tool that automatically writes the results into temp files for each call. Works quite well.
Systems of record have never been more important than in a world where you have AI agents that will do 100X more work on these platforms than people ever did.
Agents will be querying the data in these systems, processing tasks, executing workflows, collaborating with human and agent users, and more.
Thus, the governance, reliability, security, access controls, and business logic remain critical more than ever — the OpenAI Hugging Face incident is just a small peek into what the future will look like with agents running around trying to execute on their goals.
But this only works for the systems of record that respond with the right product experiences, APIs, and business models to support agents on and off their platform executing in these systems.
Huge moment of opportunity and disruption right now in all of software.
@SamBroner Instead, SaaS should be giving you a full harness tailored exactly to their underlying system which you have full control over and can connect to it however you like.