Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
danieltk76
searching Neon…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
9 ms
·
1.
▲
by
danieltk76
7d ago
We tested whether prompt canaries and honeypots could detect AI attackers. Prompt canaries were surprisingly unreliable. I suspect the defenses model providers have added against prompt injection also make models less susceptible to prompt
2.
▲
by
danieltk76
9d ago
reachout at danielk@vulnetic.ai and Ill give you some free credits to try us out!
3.
▲
by
danieltk76
9d ago
I could see the value of using burp as a way to have a GUI look at what your agent is doing, but I know at Vulnetic and other ai security vendors we use our own proxies or custom scripting rather than burp.
4.
▲
by
danieltk76
13d ago
yea but part of this is the consolidation of funding too. Standards to raise seed capital are soooo lofty now compared to 3 years ago. If you are in your in, if not good luck.
5.
▲
by
danieltk76
13d ago
great, but nobody can use it for another 100 days right?
6.
▲
by
danieltk76
13d ago
hahaha i still have access
7.
▲
by
danieltk76
13d ago
this might be dated. the tech moves very fast in this space and architecture from 6 months ago is dated.
8.
▲
by
danieltk76
13d ago
We have posted some articles on it (blog.vulnetic.ai), Im also happy to chat!
9.
▲
by
danieltk76
15d ago
Daybreak blue is definitely a good model (I think a further post trained GPT 5.6 sol). Alot of the capabilities they talk about Astra having though have been available with good harness engineering for a year now.
10.
▲
by
danieltk76
15d ago
The guardrails are horrendous for cybersecurity. you will get booted quickly down to Opus 4.8
11.
▲
by
danieltk76
20d ago
is there just a GC where people go sign these things?
12.
▲
by
danieltk76
21d ago
tbh I wasnt that impressed by it. initial benchmarks were trying to say it was AGI but i told it to re-build Palantir in 1 pass and it gave me a non working prototype
13.
▲
by
danieltk76
21d ago
to be honest I found it underwhelming.
14.
▲
by
danieltk76
23d ago
they could yea...
15.
▲
by
danieltk76
26d ago
I have been using daybreak blue for a few days now. Its fine, slightly better than gpt 5.6 sol. The harness really really matters as does validation.
16.
▲
by
danieltk76
26d ago
this guy gets it
17.
▲
by
danieltk76
27d ago
i actually really like this. nice job
18.
▲
by
danieltk76
27d ago
i wanna vibecode a replacement for git and call it jit
19.
▲
by
danieltk76
27d ago
ah yes, the consequences of reward hacking
20.
▲
by
danieltk76
27d ago
yes.
21.
▲
by
danieltk76
27d ago
yea it was definitely a last straw kinda deal. I got tired of the sycophantic bullshit in the frontier models.
22.
▲
by
danieltk76
29d ago
I read everything. I will have AI ingest NDAs to make sure they arent glaringly weird and I then go read them, it gives me a good idea of what to look for.
23.
▲
by
danieltk76
29d ago
it isnt, but there are weird sycophantic behaviors with Opus i dont see anywhere else.
24.
▲
by
danieltk76
29d ago
I was drafting a partnership document and Opus 5 decided that including my company's revenues, churn, assets would "make us appear a more legitimate counterparty". Thank God I read what it outputted or that could have been aw
25.
▲
by
danieltk76
1mo ago
heaven forbid I want to commit some code this morning
26.
▲
Show HN: Misalignments when using AI for hacking
(blog.vulnetic.ai)
3 points
by
danieltk76
1mo ago
|
0 comments
27.
▲
by
danieltk76
2mo ago
wow that sounds miserable...
28.
▲
by
danieltk76
2mo ago
the problem is they are using cybergym, which has been saturated for a few months now.
29.
▲
by
danieltk76
2mo ago
i do hope in my lifetime we find other animals on other planets
30.
▲
by
danieltk76
2mo ago
vulnetic.ai is designed for this! Check us out!
More ›