_pdp_

3 ms·

31.
by _pdp_ 1mo ago
All computer primitives are relatively straightforward in pure form and vastly more complicated in real-world scenarios.
32.
by _pdp_ 1mo ago
> we are just a couple of steps from the possibility of a true rogue AI No no. We are not a couple of steps away. This is happening. AI is already used for hacking and creating a harness that makes this fully autonomous is relatively str
33.
by _pdp_ 1mo ago
detect, slice & dice, re-prompt
34.
by _pdp_ 1mo ago
Doom loops are as much of a model problem as it is deficiency of the harness. I have not seen any other open source harness that deals with them except the one I started because of this obvious gap. See my other comment with examples where
35.
by _pdp_ 1mo ago
Ox Alpha has been running on auto-pilot for the past 5 days on various experiments. Very impressive model. Here are some examples, open-source documented and the data available in HF datasets: https://openzot.github.io/whets
36.
by _pdp_ 1mo ago
A harness is all that plus compensating for bad model behaviour - which unfortunately few harnesses actually do
37.
by _pdp_ 1mo ago
I've seen a lot of comments here and elsewhere about why is a proxy worth $8b. It is a bit more than a proxy. It is a known brand for model routing and it has developed means to route model on their own and 3rd-party infrastructure at
38.
by _pdp_ 1mo ago
It may come across harsh but IMHO the only interesting thing about this project is that it is written in Zig. That's it. Everything else in the harness is largely the same just Vercel-flavoured. The portability benefit is also a bit ov
39.
OpenAI Daybreak Red(developers.openai.com)
2 points by _pdp_ 2mo ago|0 comments
40.
by _pdp_ 2mo ago
Considering that ripgrep, git, and, you know, other dev tools are part of the toolbox, then why ship them inside this executable? And, furthermore, if you ship them, then why stop there?
41.
by _pdp_ 2mo ago
Plenty of previous art.
42.
by _pdp_ 2mo ago
Very interesting. Thanks for sharing.
43.
by _pdp_ 2mo ago
If you can spin up a new software project with little to no effort, the most important part of the job becomes making the right decisions and this is where we will be spending 99.9% of your time.
44.
by _pdp_ 2mo ago
Full disclosure I'm a founder of cbk.ai and chatbotkit.com, where we've built reusable agent components. I have never really saw flowise, n8n the now defunct OpenAI agent builder and other similar products as competitors and I wan
45.
by _pdp_ 2mo ago
Because it is a workflow builder. It is the wrong mental model. Agents and workflows are completely different things.
46.
by _pdp_ 2mo ago
I might restart my old security blog then... anyway I wonder what will happen. I think it might get flooded by automated AI submissions.
47.
Security Incident INC-2026-07-28-01 – UK AI Security Institute [pdf](cdn.prod.website-files.com)
65 points by _pdp_ 2mo ago|54 comments
48.
by _pdp_ 2mo ago
How does that help 99.9% of all other companies though?
49.
by _pdp_ 2mo ago
Wake me when they deliver Opus 4.8 level performance for $5 per million tokens.
50.
by _pdp_ 2mo ago
I mean yah... host glm and kimi and I am game.
51.
by _pdp_ 2mo ago
I have mixed feelings about software factories! On one hand, our core product is just simply not fit for them at its scale. We've tried but the project is large enough to require human input for every change. But we have AI automations
52.
by _pdp_ 2mo ago
Part of the threat model. If that is an issue then don't add the capability to the agent. It is simple as that. But we cannot guarantee that the agent wont leak some information. So if you connect personal accounts (i.e. oauth) this in
53.
by _pdp_ 2mo ago
I don't think there is any dispute there is a real risk. But hype does not really help shape the conversation and this is the problem. I am sure both companies know more than they can disclose and that gives them unique perspective out
54.
by _pdp_ 2mo ago
I am not saying it is marketing but typically when there is a data breach you may hear from the CISO but most of the time is is vague PR response. In this case I get loud signals from both HG and OpenAI leadership without much information e
55.
by _pdp_ 2mo ago
Perhaps you are right. We've only been using them for two weeks.
56.
by _pdp_ 2mo ago
Hang on a second. You think accessing private information in a public channel is a good idea? If a tool is authenticated as personal then it can be only accessed under personal context - i.e. privately. Now if the author of the agent wants
57.
by _pdp_ 2mo ago
We've tried them. So far quite expensive with not much return. This channel might not work as well as search.
58.
by _pdp_ 2mo ago
No channel will disclose private information. The only way to get that to work is to DM the agent directly or to use slash commands which are ephemeral / private.
59.
by _pdp_ 2mo ago
It is safe bet to assume that all of them are generated with LLMs.
60.
by _pdp_ 2mo ago
I agree with what you are saying but this is not how we solved the issue with who can access what. In our cloud harness we have the notion of shared and private secrets. When something is authenticated with a shared secret it is basically s

More ›