Absolutely delusional. Guy doesn't even work for you and you're wanting the candidate to answer the phone at an ungodly hour and drop everything for what could be a dud.
@victorianoi TIENES TODA LA RAZÓN LE VOY A LLAMAR AHORA POR LA NOCHE A LAS 23:28 que seguro que entenderá que los que montamos Startups llamamos a cualquier hora: ES EL PERFIL!!!
This question is asked at Jane Street for $750k/year roles:
> A drunk man stands 1 step from a cliff edge.
> Every second he randomly steps forward or back with 50/50 odds.
> What's the probability he eventually falls off?
if you answered 50%, you just failed.
drop your answer below. I'll reply to the right ones.
Symphony take 2.
This is basically Symphony by @alex_frantic but with the DAG of tasks created (maybe adjusted as well??) dynamically in house instead of in @linear.
This might be a take that finally works in a token efficient way!
@_lopopolo try it out!
Claude Code is about to release a feature called /workflows that I think will be extremely significant.
Especially for Enterprise AI.
I talked about this in 2024 in a post called Companies Are Just Graphs of Algorithms.
Basically the idea is that all work is just an algorithm, i.e., a series of steps to accomplish a goal.
Skills and Cowork have been heading in this direction already, and we've seen what that's done to company valuations in various spaces.
Well this is closer to the final form.
It's turning the regular, expected work that's done in companies into pseudo-deterministic workflows that follow defined SOPs.
The human role will be determining what problems to solve (taste, expeirence, etc), building new products from that, and then optimizing these workflows from above.
But the work itself will be these workflows executed according to SOPs.
How should multi-step agent orchestration look like, maybe something like this.
* Discrete steps each a separate context window
* Structured outputs as handoffs between steps
* Evals on every important operation
* Routing based on eval steps
can someone PLEASE launch OS claude managed agents??
I’ll pay anything. Was building this myself at some point but would prefer someone do it with 100% focus.
generational opportunity right here. i can list 30-50 companies that would use it instantly
@michael_chomsky I meant, it's unclear to me what you mean by OS within the context of CMA, though to be fair I haven't looked at that API yet. I'm building an orchestration layer using the claude / codex binaries as the inference side
@alexndrxc@zuess05 Yes, and that's the more pernicious one, in order to guide the architecture of a codebase you need to be able to understand how it works internally, even with the best steering they'll still produce slop to varying degrees
@vaibcode not only that, the tools we use to understand the scope of the code changes needs to change as well, we need better surfaces than the current PR diff view