[bitmasher9]

I think we’ll start seeing consulting agencies advertise how many vulnerabilities that can resolve per million token, and engineering teams feeling pressure to merge this generated code.

We’ll also see more token heavy services like dependabot, sonar cube, etc that specialize in providing security related PR Reviews and codebase audits.

This is one of the spaces where a small team could build something that quickly pulls great ARR numbers.

[contractlens_hn]

The same vertical-specialist logic applies in legal tech. Law firms are drowning in contract review — NDA, MSAs, leases — and generic AI gives them vague answers with no accountability. The teams winning there aren't building 'AI for lawyers', they're building AI that cites every answer to a specific clause and pins professional liability to the output. That's a very different product than a chatbot.

[dgb23]

What is needed there are custom harnesses that don’t let the LLM decide what to do when. Use their power of pattern matching on data, not on decision transcriptions.

[delecti]

Does SonarCube use LLMs these days? It always seemed like a bloated, Goodhart's law inviting, waste of time, so hearing that doesn't surprise me at all.