@gregkh I've noticed a lot of so-called AI startups have products which are basically "give us read access to all your code" and it will be like 2 Chinese, Indian or Russian male names plus an LLM as the ostensible authors. oh heck no....
@gregkh That's ... a lot. A lot "lack of thought given", I mean.
I'm pretty sure even their own tool would inform them about many ways this is not great.
(Advanced pattern recognition may even be actually useful in security engineering. But, uh, I'd prefer that to not come from the oligarch-kleptocracy.)
But many so-called FOSS supporters also love vibe coding, and proud to share their AI-generated spaghetti code.
@gregkh
@gregkh @gunchleoc we don't have closed source. They already train on all our code...
@gregkh something like half of all the AI startups have to exit by end of 2026 - to break even or make a little money - the days of infinite cash ended in 2025 ... the big entities will rationalise a new sector has been made. They will not fail but stock will be corrected. After all this we can have saner conversations.
@gregkh "free during the evaluation periods" so you gotta put in a credit card? 🤣
👉 "a handpicked group of non-commercial Open Source maintainers"
👉 looks inside
👉 "What is your company's full, legal name?"
Great job, guys. 🤦
@gregkh I bet this is a clever way to avoid the pain and suffering of coordinating vulnerabilities while still wanting the attention
LLMs can find vulnerabilities, but then it takes enormous efforts to patch and coordinate the vulnerability
Unless you Tom Sawyer the project into doing it for you
@gregkh already!!!
Can I post this on one of the other platforms? Not sure which yet. LinkedIn?