Recent submissions

Comment · 4 months ago
Comment · 4 months ago
This looks really interesting. Optimizing for humans vs. agents feels like the new wave of Desktop vs. Mobile (where mobile won) - agents are going to win even faster.

Where is the repo? It's mentioned but I can't find it.

Comment · 8 months ago
That is crazy. You are suggesting that corporations should have no power over their own IP.

Are you really saying that if Anthropic sells a limited version of their product to Palantir at a certain price, the government should be able to demand access to an unlimited version of Anthropic's product for free because they are a customer of Palantir?

That would effectively mean the government gets an unlimited license to all IP of companies that do business with government suppliers... that would be terrible.

Comment · 11 months ago
Love seeing AI applied to something actually painful: deciphering PG&E’s criminally convoluted bills. If this really works as claimed, it's a public service.
Comment · 1 year ago
> That's the job of each employee, CTO, and architects, together. Not the CEO. The CEO shouldn't make technical decisions like that.

Using your company's product isn't a technical decision. This is a ridiculous statement.

> OTOH, I remember all the toxic managers (along with the HR, CEO, and other spineless people) who told me to chill out after having verbally abused people

Why are you verbally abusing people? That is not very nice. I agree that you should chill out and stop doing that.

Comment · 1 year ago
Sounds like you come with some baggage. Maybe take a break from work for a while - you must not enjoy it.

We build a zero trust AI layer for companies... this is literally a demand that we dogfood our own product.

How dare a CEO ask employees to use their own product inside a company? That's a terrible idea and never leads to anything good.

Comment · 1 year ago
If you are building or using AI agents, three critical lessons I learned:

1. There is no "best" model or tool 2. Build for today's ecosystem (stdin/stdout), not tomorrow's (SSE) 3. MCP is stateful — and that is going to change *everything*

Comment · 1 year ago
Odd timing - right in the middle of RSA in SF. Llama (and other US-trained open weight models) are key to national security and cybersecurity, and there's a built in audience 30 minutes away if this were two days later or two days earlier...

Live from the official Hacker News API

A modern reader for Hacker News · SvelteKit + Skeleton