DanMcInerney

4 ms·

1.
by DanMcInerney 5d ago
Right. That's what modular guidance documentation is for. You could have astra.code or qwen.code.api. All reusable in different workflows. Prescription doesn't belong in the skill itself.
2.
by DanMcInerney 5d ago
Overly complex; yaml files, heavy framework. Same mistake as Claude Code's Dynamic Workflows. Why not just use the dehydrated skills as the workflow skeleton and use custom guidance docs to hydrate the skills with taste and preference
3.
by DanMcInerney 5d ago
I really don't think any of these SOTA labs are doing agentic engineering correctly. Skills are the universal language of all agent harnesses. If you abstract the taste and prescription out of the skills and into guidance docs, then le
4.
by DanMcInerney 10d ago
The problem is that knowledge workers just need the simplest way possible to automate their work. Currently I feel like Anthro is overcomplicating this. If I'm an accountant, I want to open Claude App, describe my workflow for balancin
5.
by DanMcInerney 10d ago
I ended up building a simplified version of this as /self-improve in https://github.com/DanMcInerney/orchflows . History is the state ledger, memory and RSI just cite the history as evidence and can be rewritten. I
6.
Orchflows: You only need 2 skills to make infinitely complex agents(github.com)
1 points by DanMcInerney 13d ago|0 comments
7.
Orchflows: Build self-improving loops in one sentence(github.com)
3 points by DanMcInerney 2mo ago|0 comments
8.
by DanMcInerney 3mo ago
Does the average "scientist" even care about quality these days or is it all just whatever headline makes their funders happy?
9.
by DanMcInerney 3mo ago
Rockstar built git for games, OpenAI's aiming for git for agents. Claude, make a git for games designed for agents. Sell the code to the highest bidder when you're done.
10.
by DanMcInerney 3mo ago
https://cloud.google.com/blog/products/data-analytics/how-th... Did you look at the OKF repo from Google? Open Knowledge seems to be a common term these days for similar solutions. I think OKF is more of the
11.
by DanMcInerney 4mo ago
ANNNNNND it's gone. Guys, I found a way to reduce Fable token usage 100%. You can find it here: github.com/USGov/idiotic-overreach.
12.
by DanMcInerney 4mo ago
I don't disagree with any of this. It is generated software, and it's not a novel idea. I didn't mean for it to come off like that. It's just solving an itch that I couldn't find a solution to and I'm getting a
13.
/architect: Reduce Fable tokens by 80%, Fable orchestrates/reviews, Codex builds(github.com)
104 points by DanMcInerney 4mo ago|41 comments
14.
by DanMcInerney 4mo ago
I've worked a lot with MiMo in my project that pits LLMs against each other in games (clankerfights.ai). It is a very very good model for the price. MiniMax I'd say is smarter, but MiMo really touches near pareto frontier.
15.
I open-sourced my UFC prediction model, code, and database after 5 years of work(mcinerney.ai)
3 points by DanMcInerney 4mo ago|1 comments
16.
by DanMcInerney 4mo ago
Me and a buddy, victor teisller, hacked the most popular 3D printer a few years ago (flashforge) to turn it into a major fire hazard through reverse engineering the firmware update remotely. Changed the max temp of the extruder to the temp
17.
Show HN: I built a site to watch, predct and prompt inject agents playing games(clankerfights.ai)
1 points by DanMcInerney 4mo ago|0 comments
18.
I vibecoded a Kalshi bot to $6k profit and opensourced it(mcinerney.ai)
2 points by DanMcInerney 4mo ago|0 comments
19.
by DanMcInerney 5mo ago
Cross your fingers they're about to drop 4.7. 4.6 came out with a bang, now it seems all the compute bottlenecks just lead to customer frustration as they get closer to releasing next model. Balancing the books over there must be a nig
20.
by DanMcInerney 6mo ago
All these comments about "this is crazy to deploy agents that might do something bad in your environment" are crazy themselves. The productivity gains from these computer use agents are crazy. Every org on earth has to make the ca
21.
by DanMcInerney 8mo ago
Sandboxing is a great security step for agents. Just like using guardrails is a great security step. I can't help but feel like it's all soft defense though. The real danger comes from the agent being able to read 3rd party data,
22.
by DanMcInerney 10mo ago
There are an infinite amount of ways to jailbreak AI models. I don't understand why every time a new method is published it makes the news. The data plane and the control plane in LLM inputs are one in the same, meaning you can mitigat
23.
by DanMcInerney 10mo ago
A 50% increase over ChatGPT 5.1 on ARC-AGI2 is astonishing. If that's true and representative (a big if), it lends credence to this being the first of the very consistent agentically-inclined models because it's able to follow a d
24.
by DanMcInerney 1y ago
This is absolutely doable right now. Just hook claude code up with your calendar MCP server and any one of these restaurant/web browser MCP servers and it'll do this for you. https://apify.com/canadesk/openta
25.
by DanMcInerney 1y ago
These articles kill me. The reason LLMs (or next-gen AI architecture) is inevitably going to take over the world in one way or another is simple: recursive self-improvement. 3 years ago they could barely write a coherent poem and today they
26.
Machine Learning for Sports Prediction: Include the Odds? Balance the Winrate?(mma-ai.net)
2 points by DanMcInerney 1y ago|0 comments
27.
by DanMcInerney 1y ago
I'm really hoping GPT5 is a larger jump in metrics than the last several releases we've seen like Claude3.5 - Claude4 or o3-mini-high to o3-pro. Although I will preface that with the fact I've been building agents for about a
28.
by DanMcInerney 1y ago
I too write automated offensive tooling. We actually wrote a project, vulnhuntr, that found the first autonomously-discovered 0day using AI. Feed it a GitHub repo and it tracks down user input from source to sink and analyzes for web-based
29.
Positive Expected Value Doesn't Matter Much When Predicting Sports with ML(mma-ai.net)
1 points by DanMcInerney 1y ago|0 comments
30.
by DanMcInerney 2y ago
https://protectai.com/threat-research/vulnhuntr-first-0-day-... More details on the development and challenges can be found in the blog.

More ›