danenania

5 ms·

31.
by danenania 8mo ago
There are different kinds of tribal knowledge. Some is company-specific, some is role-specific or domain-specific.
32.
by danenania 9mo ago
The date is just a useful fiction to: - Create urgency - Keep scope creep under control - Prioritize whatever is most valuable and/or can stand on its own If you just say “I don’t know” and have no target, even if that’s more honest, t
33.
by danenania 9mo ago
I don’t think “ready to merge” necessarily means the agent actually merges. Just that it’s gone as far as it can automatically. It’s up to you whether to review at that point or merge, depending on the project and the stakes. If there are C
34.
by danenania 9mo ago
Humans make subtle errors all the time too though. AI results still need to be checked over for anything important, but it's on a vector toward being much more reliable than a human for any kind of repetitive task. Currently, if you
35.
by danenania 9mo ago
I think it’s just realpolitik grand chessboard strategy. Knocking out an unfriendly/uncooperative leader of a strategically important country. That’s always been the real justification for US foreign policy. It’s a game of risk, withou
36.
by danenania 9mo ago
Oh man, I relate so hard on the sports conversations.
37.
by danenania 9mo ago
I definitely see your point. I'd just say though that it can put a lot of pressure on the romantic relationship. Some can handle it; others might not. And also it makes it much more difficult to recover if things don't work out. S
38.
by danenania 9mo ago
Going further, you don't even need to count your reps or track how much weight you're lifting. Literally just do any exercise with any weight per muscle group to near failure for 2-5 sets. Rest the muscle groups you targeted the n
39.
by danenania 10mo ago
That’s because gods are a mythical/supernatural invention. No technology can ever really be omniscient or omnipotent. It will always have limitations. In reality, even an ASI won’t know your intent unless you communicate it clearly and
40.
by danenania 10mo ago
Thanks for the comment. - On precision vs. noise: yeah, this is a core challenge. Quick answer is the scanner tries to be conservative and lean towards not flagging borderline issues. There's a custom guidance field in the config that
41.
by danenania 10mo ago
Hey all, I'm an engineer at Promptfoo (open source evals and red teaming for AI). We're launching a tool that scans GitHub PRs for common LLM-related vulnerabilities. This post goes into detail on how it was built and the kinds of
42.
Building a Security Scanner for LLM Apps(promptfoo.dev)
7 points by danenania 10mo ago|3 comments
43.
by danenania 10mo ago
You can use an agent while still understanding the code it generates in detail. In high stakes areas, I go through it line by line and symbol by symbol. And I rarely accept the first attempt. It’s not very different from continually refinin
44.
by danenania 10mo ago
I often use multiple documents to plan things that are too large to fit into a single planning mode session. It works great. You can also use it in conjunction with planning mode—use the documents to pin everything down at a high-to-medium
45.
by danenania 10mo ago
A lot of that would be people working on proprietary code I guess. And most of the people I know who are doing this are building stuff, not streaming or making videos. But I'm sure there must be content out there—none of this is a secr
46.
by danenania 10mo ago
I know of many experienced and capable engineers working on complex stuff who are driving basically all their development through agents. This includes production level work. This is the norm now in the SV startup world at least. You don&#x
47.
by danenania 10mo ago
A problem I’ve found with LLMs for docs is that they are like ten times too wordy. They want to document every path and edge case rather focusing on what really matters. It can be addressed with prompting, but you have to fight this const
48.
by danenania 10mo ago
Agreed, how dare they break the law like that!
49.
by danenania 10mo ago
Non-developers usually prefer them to IDE or terminal based tools.
50.
by danenania 10mo ago
Investors are really bad at predicting up front what can become a unicorn and what can’t.
51.
by danenania 11mo ago
A question is which side agents will achieve human-level skill at first. It wouldn’t surprise me if doing the work itself end-to-end (to a market-ready standard) remains in the uncanny valley for quite some time, while “fuzzier” roles like
52.
by danenania 11mo ago
Though I think the CEO role is realistically one of the hardest to automate, I’d say middle management is a very juicy target. To the extent a manager is just organizing and coordinating rather than setting strategic direction, I think that
53.
by danenania 11mo ago
They’re also great, in my experience, for overcoming writer’s block and procrastination. Just as a rubber duck to bounce ideas off of and follow different threads. It makes the writing process faster and more enjoyable, despite never using
54.
by danenania 11mo ago
It’s because the LLM is a commodity. What differentiates a product is not the commodity layer it’s built on (databases, programming languages, open source libraries, OS apis, hosting, etc) but how it all gets glued together into something u
55.
by danenania 11mo ago
The post is about an attack carried out with the help of claude code, a coding agent.
56.
by danenania 11mo ago
I think that makes sense. The change is not really in the kind of attack—anything the agent can do a human attacker could also do—but in the amount of effort and expertise required to design and scale up the attack. It’s a quantitative ra
57.
by danenania 11mo ago
If you need this, could you just put your own kv cache in front?
58.
by danenania 11mo ago
It is definitely a buzzword, but agents also are legitimately changing many fundamental things about security, so…
59.
by danenania 11mo ago
I wrote this for my company’s blog about the recent hack that Anthropic reported, where a China-linked group carried out an attack against many targets using Claude Code. It goes into the tension between capabilities and safety (from a secu
60.
Will agents hack everything?(promptfoo.dev)
6 points by danenania 11mo ago|9 comments

More ›