ramoz

18 ms·

361.
by ramoz 10mo ago
I think we need to shift our mindset on what an agent is. The LLM is a brain in a vat connected far away. The agent sits on your device, as a mech suit for that brain, and can pretty much do damn near anything on that machine. It's the
362.
by ramoz 10mo ago
You are missing an entire agentic experience. And I wouldn't call it vibe coding for an engineer; you're more or less empowered to truly orchestrate the development of your system. Cursor has agent, but that's like whoever el
363.
by ramoz 10mo ago
What he meant was, agents will probably not be these web abstractions that run in deployed services (langchain, crew); agents meaning the Harnesses (software wrapper) specifically that call the LLM API. It runs on your computer because of i
364.
The Model T of Agentic AI was Introduced in 2025(backnotprop.substack.com)
1 points by ramoz 10mo ago|0 comments
365.
Show HN: Debug Console for Claude Code(github.com)
1 points by ramoz 10mo ago|0 comments
366.
by ramoz 10mo ago
Claude Code hooks is a great way to integrate these things
367.
by ramoz 10mo ago
Send them this https://backnotprop.substack.com/p/50-first-dates-with-mr-me...
368.
Show HN: Cupcake – Better performance and security for coding agents (via OPA)(github.com)
12 points by ramoz 10mo ago|1 comments
369.
50 First Dates with Mr. Meeseeks(backnotprop.substack.com)
4 points by ramoz 11mo ago|0 comments
370.
Who Owns Alignment?(backnotprop.substack.com)
2 points by ramoz 11mo ago|0 comments
371.
by ramoz 11mo ago
No company has come close to producing as close as of a reliable agentic coding solution as Anthropic has.
372.
by ramoz 11mo ago
title fixed per microsoft correction
373.
by ramoz 11mo ago
Coding - e.g. Claude Code, Cursor both announced 1B revenue run rates.
374.
by ramoz 11mo ago
I meet with enterprise clients who explore things like Copilot Studio. Microsoft platforms move too slowly too keep up with innovation pace, and suffer from classic platform restriction in regards to building useful, relevant, and *reliable
375.
by ramoz 11mo ago
Github Copilot review is such a simple - optional - ux, it makes sense as a feature. I enjoy it.
376.
Microsoft lowers AI software growth targets(finance.yahoo.com)
130 points by ramoz 11mo ago|97 comments
377.
by ramoz 11mo ago
It'll be a common paradigm. Some agents support the coding agent discover relevant context for a plan, others will help the agent stay on track and ensure no rules break.
378.
by ramoz 11mo ago
I wrote the comment while waiting on a 50min rust build pipeline
379.
by ramoz 11mo ago
Rules need empowerment. Excited to be releasing cupcake at the end of this week. For deterministic and non-deterministic guardrailing. It integrates via hooks (we created the feature request to anthropic for Claude code). https:/&#x2F
380.
by ramoz 11mo ago
At the end of this week we are releasing https://github.com/eqtylab/cupcake You can see all of Claude’s commits. I’ve shipped so much with ai. Favorite has been metrics dashboards of various kinds - across life and bus
381.
by ramoz 11mo ago
The answer is typescript is a much simpler and more pleasant developer experience than any other language. These are products they need to, and often originate from, fast churn of code/features. Otherwise they’d be building these types
382.
by ramoz 11mo ago
Claude Code moved to partial file reads over the summer. Super premature optimization. It’ll hallucinate what lines it needs to read, it’ll continuously miss critical context in favor of trimming tokens. Luckily we can now hook and force th
383.
Does an A100 still make profit(a100economics.com)
3 points by ramoz 11mo ago|1 comments
384.
by ramoz 11mo ago
context: https://x.com/burrytracker/status/1991289685474439579?s=20
385.
by ramoz 11mo ago
You wouldnt feed it numerical data, but you would allow it to make certain calculations (via tools of a harness) as it relates to your portfolio.
386.
by ramoz 11mo ago
Ultimately, relying on system level instructions is unreliable over time. Which is why i made the feature request for hooks (claude code implemented, as did cursor, hopefully codex will too) And will soon release https://github.c
387.
by ramoz 11mo ago
After creating the feature request for Claude Code hooks[1] a few months back, Cupcake is nearly ready for release. Cupcake is a governance/policy-enforcement layer for agents. Its innovation is binding OPA/rego to agent runtimes
388.
by ramoz 1y ago
EQTY Lab | Software Engineers (Fullstack / AI ), or DevRel | SF, LA (looking for local, but wfh if cool) | Full Time | https://www.eqtylab.io/ We are hiring software engineer(s) to help innovate on applied agentic appl
389.
by ramoz 1y ago
Hooks are underutilized and will be critical for long-running and better agent performance. Excited to release Cupcake in the coming weeks. What I started building when i made the feature request for hooks. - https://github.com&#
390.
by ramoz 1y ago
is this anecdotal?

More ›