Alex

Building @mato_hq - AI agents that interview humans.
Built Blikket ($150M+ in client revenue). Coding since 11. Norwegian in Los Angeles.

[Alex @abenz95]

I'm building @mato_hq, AI agents that interview humans, starting with podcasts.

$0K to $18.5K MRR in 3 months; nobody has churned; about 30 shows running. one of them sits at #10 in News Commentary on Apple Podcasts.

before this, I spent years turning other companies' conversion rates into revenue, $150 million in added revenue to be exact, wrote a book about it, and got a wireframe method design-registered in Norway twice. now I do it for my own company.

bootstrapped the journey so far.

[Alex @abenz95]

@Austen Benchmarks keep testing the model in isolation. The real gap is the harness around it: retries, receipts, and knowing when the agent should stop.

[Alex @abenz95]

Clip tools keep getting sold as the thing that makes a podcast travel. they don't make anything. they find a loud 40 seconds and crop it. the moment was already made or not made when somebody asked the guest one more question instead of moving on. you can't crop your way to that.

[Alex @abenz95]

@benjitaylor the gap between a good designer and a bad one has never been more obvious tho

[Alex @abenz95]

uploading my coding sessions to paxel for @ycombinator F26 application. it seems to struggle with the amount of sessions I have haha

[Alex @abenz95]

just finished a new interview about @mato_hq and @ycombinator. go check it out!

https://t.co/xXr7cXvx2d

[Alex @abenz95]

@DozenDucc @yiding_song @theWaddleLabs 20 minutes is wild. what happens when the robot hits a physical edge case the agent didn't anticipate, does it re-prompt or just stop?

[Alex @abenz95]

@openclaw naming NVIDIA as the security partner when your threat model eventually includes NVIDIA hardware seems like a thing worth asking before the press release

[Alex @abenz95]

@omarvvvr what's the excuse when codex writes 80% of it for you

[Alex @abenz95]

did they change opus 5 to 200k context window?

[Alex @abenz95]

@cognition curious how it holds on multi-file refactors, that's where frontier models actually separate from open source in practice

[Alex @abenz95]

does opus 5 keep doing this with you too? it wants to pick up tomorrow almost always.

does it have plans I’m not aware of?

[Alex @abenz95]

@jaltucher 1.3M tokens and 10 hours to ship a playable game tho, curious how the second session holds up

[Alex @abenz95]

@jonbeckman Grok 4.5 is actually pretty brutal at calling out overengineering in my experience, might be the right fit for what you're trying to defend against

[Alex @abenz95]

Turning one episode into 40 posts is the AI pitch everyone's running right now. Forty outfits, one thin idea underneath.

If the episode had a line worth stealing, you never needed the pipeline. If it didn't, the pipeline just spreads the boredom wider.

What's the one line from your last episode someone would actually repeat?

[Alex @abenz95]

@eve_bouff @ycombinator @snowmaker @bradflora this is too funny, need to try this!

[Alex @abenz95]

anyone else having issues with opus 500 error?

[Alex @abenz95]

@tomdoes_tech good luck with codex, they nerfed all limits this week. I’m on the $200 plan and used all my usage in 1 day

[Alex @abenz95]

opus 5 is a beast

[Alex @abenz95]

@githubstatus something going on with PRs?