FWIW, this Nobel Laureate (Hinton) disagrees with you about consciousness. Maybe you should be less certain about your credences.

Anyway, there was some discussion about the Goblin Problem and its relation to consciousness it in the latest Last Week in AI. Always worth a listen.

Remember Solyndra?

Jesus, this is a brain-dead take. Anthropic is the AFAICT only for-profit company that takes AI safety and alignment seriously.

Dario (and others') work on Constitutional AI is AFAICT the only realistic solution to this very real engineering problem. And they publish what could be their secret-sauce constitution, verbatim, under a Creative Commons license.

This. And note the headlines for what afaict is the same product. For Meta they are enabling sexual predators. For Apple they are ... making style and design choices?

As a person who worked mightily hard for over a year on a cross functional team with top management attention to bring a product into official FedRAMP compliance... for all the blood sweat weekends late nights documenting automating improving processes and tooling... FUCK YOU Microsoft for cheating your way in and endangering your country and your customers in the process. FUCK. YOU.

The Black Blade was forged a million billion years ago. Its cosmic story goes on for eternity. It master is its slave. Hahahahahaaaaaa

LibreWolf is FF with the enshittification removed

Yup. 'Orchestration' is the fancy word. But it's just "scripting with LLMs and fancy Python packages"

I see a similar trend inside the Big Software company I work at. We're automating fixing the bug backlog, automating test fixes, automating platform posting, ... I don't think anybody knows how this all shakes out but we know it's the future. Strong 1990's Internet vibe.

When I see posts about how crappy AI is, it's not usually from people who are actually using it to get real work done.

In the current state of flux in both tech, and global geopolitics, I would not bet on *any* tech company being around in 100 years. No matter how big they are RN.

It does not take many bold but harebrained plans (could cough Neom Line) to drain even a trillion-dollar fund.

9 out of 14... with a sample that small, this is essentially indistinguishable random chance

No one thinks AI is that great

You have not been hanging around computational biologists lately. I can tell.

Seems pretty stupid to double down on pissing off people who are effectively godlike wizards who can eat your entire lifes data with ketchup before breakfast

Story by p k dick is what makes it so awesome

He means Unicode obvs