Michael Crowley

CS student @ Trinity College Dublin. Co-founder and CTO of Lockup

[Michael Crowley @notmcrowley]

My website, https://t.co/KhIlAzbCwB just reached the top 10 posts on Hacker News! This means a lot to me as a first year CS student and I really appreciate all the feedback I have received since yesterday.

[Michael Crowley @notmcrowley]

Lockup 2

[Ryan Morrissey @ryan_morr]

I was larping so hard three years ago damn

As a person into CS it’s a canonical event to go through a script kiddie phase

Had a flipper 0 and everything aswell

[Michael Crowley @notmcrowley]

@adelindiac_ Just get Claude Code to ssh into whatever you're hosting your agent on and tell it to get to work. (Might not be a good strategy if you are particularly fond of the files on that machine)

It set up MichaelAI to track my heartbeat in only one prompt though 🤷‍♂️

[Michael Crowley @notmcrowley]

Using Claude Code to setup Hermes or Openclaw remotely is a game changer. Your time is worth the tokens.

[Michael Crowley @notmcrowley]

@ryan_morr @hoplite_sh What harness do you guys use for the agents or is it proprietary/top secret?

[Sam Altman @sama]

so far at least, i'm pretty sure AI has been net job-creating.

this was not what i expected--although i was much less pessimistic than others, i thought by this level of capability we'd have seen some impact.

it is possible this direction keeps going!

[Claude @claudeai]

There’s hope in hard questions.

[Michael Crowley @notmcrowley]

The older the castle the more successful the startup will be.

[Rémi 🌳 @remilouf]

This is the best place to build your startup and nothing comes even close

[Michael Crowley @notmcrowley]

@ilyasut The website itself is still a work in progress. It is largely based off a list generously provided by @andrew_n_carr on here. The code is open sourced on Github and I am always open to contributions.

[Michael Crowley @notmcrowley]

All CS students should read at least a few AI papers. I started this year. It can be hard to find where to start but I mostly used a list made by @ilyasut . For this reason, I built https://t.co/KhIlAzbCwB which has all of these with some resources to make learning them easier.

[Michael Crowley @notmcrowley]

@ilyasut @scottaaronson (5/5) I’d like to end with AlexNet, the start of modern deep learning (co-authored by Sutskever). It showed CNNs could dominate image recognition, sparking a deep learning revolution

I can’t host it due to copyright, but the paper is worth reading:

https://t.co/XYfvpHaUj8

[Michael Crowley @notmcrowley]

@ilyasut (4/5) One underrated paper that I loved is Scott Aaronson’s The Coffee Automaton. It explores how complexity emerges, persists, and fades as closed systems approach equilibrium. A big thanks to @scottaaronson for letting me host this and a blog of his.

https://t.co/tJr0p2xFpd

[Michael Crowley @notmcrowley]

One thing I find really interesting while reading AI papers is how regularly the authors will reference some concept from completely unrelated fields.

I don't know if the main takeaway is how much these researchers know or how little I do at this point.

[John Collison @collision]

Ireland has ended up a vetocracy.

[Michael Crowley @notmcrowley]

@manny_karibiye @adelindiac_ Seeing these guys in the airport made me glad I'm not in competition with Call Crew.

[Dario Amodei @DarioAmodei]

Today I'm publishing a new essay, Policy on the AI Exponential. AI is progressing extremely fast—much faster than the policy process was built to handle. The essay lays out where I think the technology is now, and the action needed to close the gap: https://t.co/Lh6PWae178