Tomorrow Corporation tech demo 👏👏👏, thx James. A striking reminder of what’s possible for a dev team in control of its “tech stack”, and of how much control we sacrifice when we layer frameworks upon dependencies upon toolkits that look like free lunch, but come at a hefty { weight / performance, flexibility, maintenance } cost.
Washed Out: The Hardest Part, “the first official commissioned music video collaboration between a music artist and filmmaker made with OpenAI’s Sora video model”.
500 mile email, “a curated list of absurd software bug stories”.
AI
LLMs are like a trained circus bear that can make you porridge in your kitchen. It’s a miracle that it’s able to do it at all, but watch out because no matter how well they can act like a human on some tasks, they’re still a wild animal. They might ransack your kitchen, and they could kill you, accidentally or intentionally! — Alex Komoroske, via Simon Willison.
People seem to be falling for two rather thoughtless extremes:
“LLMs are AGI, they work like the human brain, they can reason, etc.”
“LLMs are dumb and useless.”
The reality is that LLMs are not AGI — they're a big curve fit to a very large dataset. They work via memorization and interpolation. But that interpolative curve can be tremendously useful, if you want to automate a known task that’s a match for its training data distribution.
Memorization works, as long as you don’t need to adapt to novelty. You don’t need intelligence to achieve usefulness across a set of known, fixed scenarios.
In fact, that’s the entire story of the field of AI so far: achieve increasing levels of usefulness and automation, while bypassing the problem of creating intelligence.
NeverKnowsBest: Lies, hate and the story of Emil Pagliarulo. A dissection of exponential spiralling hatefarming & defamation on youtube / reddit / twitter. Reminiscent of Hbomberguy’s video about plagiarism, but focused on { personal attacks, spitballing, entitlement, power tripping } instead of simple plagiarism. A reminder that angry liars can be dangerous when they have a platform.
