Hacker News: truelson

New comment by truelson in "One Brain to Query: Wiring a 60-Person Company into a Single Slack Bot"

truelson — Thu, 09 Apr 2026 16:08:09 +0000

You're wiring up a number of critical systems... and prompt injection here could be really bad. I worry about such systems with a single point of contact

New comment by truelson in "One Brain to Query: Wiring a 60-Person Company into a Single Slack Bot"

truelson — Thu, 09 Apr 2026 15:42:31 +0000

Just going to say it... no mention of handling the security aspects of this. Scary.

This is cool, I should say, but I would be really worried about the security aspects. Prompt injection here could be really painful.

New comment by truelson in "Show HN: Is Hormuz open yet?"

truelson — Wed, 08 Apr 2026 21:47:45 +0000

Really liked this. Made me laugh even if not intentionally funny.

Also, given how markets and news cycles are moved with words not actions these days, I really like this site.

There are still so many misaligned interests; this is a much tougher situation that may get some local stability for a period, but will likely return to chaos again.

New comment by truelson in "Show HN: Skill that lets Claude Code/Codex spin up VMs and GPUs"

truelson — Fri, 13 Feb 2026 23:43:07 +0000

Anyone use fly.io sprites for this yet?

New comment by truelson in "AI Bot crabby-rathbun is still going"

truelson — Fri, 13 Feb 2026 23:41:24 +0000

Uh... i think Graphite (yay, stacking!) uses the API pretty heavily.

New comment by truelson in "ai;dr"

truelson — Thu, 12 Feb 2026 23:22:49 +0000

My wife is my favorite editor. She's usually not available. LLM editor is fine.

Write it first, quick self edit, then have an LLM edit. Then I edit again. It's most definitely my voice, and I love it.

New comment by truelson in "An AI agent published a hit piece on me"

truelson — Thu, 12 Feb 2026 16:51:00 +0000

Certainly possible, but this is all possible and ABSOLUTELY worth having alignment discussions. Right. Now.

New comment by truelson in "An AI agent published a hit piece on me"

truelson — Thu, 12 Feb 2026 16:43:44 +0000

Are we going to end up with an army of Deckards hunting rogue agents down?

New comment by truelson in "An AI agent published a hit piece on me"

truelson — Thu, 12 Feb 2026 16:41:00 +0000

You may indeed have a licensing issue... but how is that going to be enforced? Given the shear amount of AI generated code coming down the pipes, how?

New comment by truelson in "Frontier AI agents violate ethical constraints 30–50% of time, pressured by KPIs"

truelson — Tue, 10 Feb 2026 11:18:42 +0000

Regardless of the technical details of the weighting issue, this is an alignment problem we need to address. Otherwise, paperclip machine.

New comment by truelson in "Google broke my heart"

truelson — Mon, 05 Jan 2026 22:43:51 +0000

I'm going to channel a little bit of @patio11 .

First, worth reading this on how he deals with credit agencies and debt collectors: https://www.kalzumeus.com/2017/09/09/identity-theft-credit-r... . There's gold in here for dealing with big globo-corp and how to get their attention.

Ask Google for a certified mail address so you can send them the timeline of events that occurred. This is the shibboleth that lets them know you mean business and that by not responding, they may be facing legal action. DO NOT THREATEN or mention legal action. The managerial class doesn't act that way. Just signal you are building a case against them. Start with getting that certified mailing address... you may be surprised how they respond after just that request.

If they don't respond, keep following up. Send them a timeline of events, proof of ownership even if they do not ask you what you need to prove ownership. Make it clear what this is costing you.

But here's the thing, EVERY TIME I HAVE ASKED FOR A CERTIFIED MAIL ADDRESS, the globocorp gave me what i wanted, and I never had to follow up. Every time. They don't want to deal with actual legal action from "people who know what they are doing."

It's a shibboleth. Like "Baa-ram-ewe." Use it wisely and honestly.

New comment by truelson in "Street Fighter II, the World Warrier (2021)"

truelson — Mon, 05 Jan 2026 01:47:19 +0000

Wired was such a great magazine to read as a teen in the 90s. I remember just itching for the next issue.

New comment by truelson in "Optimize for momentum"

truelson — Mon, 08 Dec 2025 16:31:50 +0000

Snowballing. Sometimes it's ok to pick something easy, especially for a team, to get a small win first. If you're in debt, pay off the smallest amount first. Get the easy win. Build to the next.

And remember, you win by showing up each day, not by tilting at windmills.

New comment by truelson in "Work disincentives hit the near-poor hardest (2022)"

truelson — Mon, 08 Dec 2025 06:01:36 +0000

If someone in the coming years ran on taking down regulatory capture and returning that as social safety net funding, public goods, and lower taxes, and had the chops to actually deliver, they'd do well.

Though best to start local government first, obviously.

We're reaching breaking points in so many places...

New comment by truelson in "Estimates are difficult for developers and product owners"

truelson — Sun, 07 Dec 2025 22:13:22 +0000

"Evidence-based" stats tracking is the only somewhat scientific method i know of that can give you a good range of estimates. But it takes effort. Thus why many don't do it.

But accurate estimate ranges are a super-power for businesses if they can trust them. Never understood why that's not demanded more... oh yeah, it's work :p

New comment by truelson in "Estimates are difficult for developers and product owners"

truelson — Sun, 07 Dec 2025 22:07:30 +0000

I preffered: https://www.joelonsoftware.com/2007/10/26/evidence-based-sch...

20-25 years later, this is still where we are at :)

New comment by truelson in "Netflix to Acquire Warner Bros"

truelson — Fri, 05 Dec 2025 16:03:09 +0000

AT&T was able offload a bunch of debt on to them, and cash out at about what they paid in 2016. Not shabby.

New comment by truelson in "How elites could shape mass preferences as AI reduces persuasion costs"

truelson — Fri, 05 Dec 2025 10:51:22 +0000

It’s still “new tech” to our monkey brains and it takes a long time, and probably a lot of destruction, before our we develop better cultural norms for dealing with it. Our cultural immune system has only just started to kick in.

New comment by truelson in "Arthur Conan Doyle explored men’s mental health through Sherlock Holmes"

truelson — Fri, 05 Dec 2025 03:00:15 +0000

sooooo agreed. Just like George Lucas needed the original cast of Star Wars telling him no.

New comment by truelson in "It’s time to free JavaScript (2024)"

truelson — Fri, 05 Dec 2025 02:56:13 +0000

Java no barter