Joanthan Mugan

Dreamed I was a sanitation worker in Hong Kong. Much more complex than I would have expected, with different trucks for different parts. I guess the night brain wants to prepare us for any eventuality.

I wrote this blog post a couple of weeks ago, but I didn’t like it, so I wrote it again. LLMs Enable Judgment: From Code to Consciousness www.jonathanmugan.com/blog/LLMs_an...

Even if large language models (LLMs) stopped improving today, they still have yet to make their real impact on computer systems. www.jonathanmugan.com/blog/LLMs_an...

I added Service Model by @aptshadow.bsky.social to my list of favorite books. www.jonathanmugan.com/FavoriteBooks/

It’s about robot goals and how interacting robots get into deadlock. It also rings true for anyone who has dealt with large bureaucracies.

The good news is that AI doesn’t want anything. We think they would take over because that is what we would do if we were them, but we were built using evolution, which rewards dominating others. They are designed, so they only want what we tell them to. I see them as glorified toasters.

It's funny, extra-compiler semantics such as comments and good variable names used to only matter for humans and not for the computer, but now with LLMs we have computers reading the code, and they do matter for those computers.

The cottage cheese that my wife brought home from the store says that it is "supernatural." I ate the whole thing. We'll see what happens.

How AI Takeover Might Happen in 2 Years — LessWrong I’m not a natural “doomsayer.” But unfortunately, part of my job as an AI safety researcher is to think about the more troubling scenarios. …

Here’s some fun reading. A well-written AI-apocalypse short story www.lesswrong.com/posts/KFJ2LF...

I feel simultaneously both over- and under-caffeinated.

The slides to my talk "What Superintelligence Will Look Like" at @datadaytexas.bsky.social are online.
www.jonathanmugan.com/WritingAndPr...

"free in-store pickup?" yeah. thats what store is.

Claude interviews Jonathan Mugan During one of our conversations, Claude asked me to share my thoughts on a particular topic. I replied that he would forget it at the end of the session, and I would save it for a day when, hopefully,...

I'm speaking Saturday at @datadaytexas.bsky.social about superintelligence. Besides talks, there's also all the coffee you can drink. I've tested it. It doesn't run out. www.linkedin.com/pulse/claude...

It's a myth that roosters wait until early morning to start crowing. They go all night. You would think that something so delicious would be more careful about giving away its position.

Claude from @anthropic.com interviews @jonathanmugan.bsky.social for his upcoming @datadaytexas.bsky.social talk:
www.linkedin.com/posts/lynnbe...

Prepare for an unforgettable journey as Jonathan Mugan weaves a tapestry of delight and inspiration. Just kidding, I'm talking about more nerd stuff.

May talk: "What Superintelligence Will Look Like" at Data Day Texas on Saturday, January 25.

You can use the discount code JONATHAN to get 20% off.

I'm here too, somewhere, ah, here @jonathanmugan.bsky.social

The future of AI is models that generate graphical interfaces. Instead of the linear, low-bandwidth metaphor of conversation, models will represent themselves to us as computers: rich visuals, direct manipulation, and instant feedback.

willwhitney.com/computing-in...

they should make eminem rap battle the guy who sings about the grinch

If you are looking for a family holiday movie, let me recommend Klaus from a few years back (on Netflix). It's funny, and like dark chocolate, the bitterness makes the sweetness more satisfying.

My next blog post as envisioned by ChatGPT

Don't understand why everyone is always worried about the fertility rate. There are too many of us. Yeah, social security is a challenge, but less of one than tearing up the planet.

When you are upset, you can’t focus on creative work. But if you can somehow focus despite the negative arousal, maybe your brain can create new ideas that you wouldn’t normally have, because it is in a different “state.” Almost like temporarily being a different person.

ChatGPT: "You're right. The approach outlined earlier doesn't differentiate between ..."
Me: "Child, I know I'm right."

Of course, if I knew everything I needed to know, I wouldn't be asking the great computer in the sky.

If you contribute to one of those academic books where each chapter is a paper and the book costs $120, aren't you essentially just hiding your paper away from the world, like in a time capsule that will never be opened?

If you work in robotics and/or AI and you like funny, read Service Model by Adrian Tchaikovsky
@aptshadow.bsky.social. It also rings true for anyone who has dealt with large bureaucracies. It’s about robot goals and how interacting robots get into deadlock.

YouTube video by demonflyingfox The Lord of the Rings - The Fellowship of the Rednecks (Official Music Video)

The internet often makes me sad, but every once in a while something comes along that makes it all worthwhile www.youtube.com/watch?v=ahq6...

An AI agent can more effectively modify its knowledge if it can read its knowledge. Maybe that’s one of the things consciousness enables for us.