AISnakeOil

I've been experimenting with AI video generation beyond raw video gen for some time now. Opus 4.5 was the first tool I tried last year, and it was okay...

Fast forward a few months, and Opus 5.5 is now running my entire video production pipeline. This model is leaps and bounds ahead of the previous model.

I can generate a complete 5-minute explainer video on my MacBook Air M4 in about five minutes from a single prompt, using the Gemini TTS API and MiniMax H3 Max for b-roll. It creates thumbnails and titles, handles the entire process, and automatically uploads the result via the YouTube API—all powered by a Claude Code plugin. (youtube.com/@ctrlaltexplain for those interested).

Why do they need to? For search, instant models are more important and fit the use case better.

Pro models are mainly for coding agent work; it doesn't necessarily make them any money.

Yeah the era of the personal computer is over.

Companies will prioritize components for cloud computing and lease us thin clients...

No it's microsoft

I guess we should be thankful that OpenAI is so open about this kind of stuff...

I'm working on a daily word puzzle game called Hinted. Wordle for people with half a brain. Streaks that are forgiving and a way to track your friend's progress. https://hintedgame.com

LLMs have awareness for the time they are spawned into memory. But it's very limited, think about if you could use your brain to think, but only after someone asked you a question. After you think the answer, then you are brain dead (unconscious) until another question is asked.

What I actually like about Apple products is the heft. They feel premium and the heaviness contributes a lot to the premium feel.

I tried a ThinkPad X1 Carbon as well, it felt like a toy.

This is how open source wins

I use to enjoy diggnation, but now every episode is about them claiming not to be alcoholics when they clearly are.

nice numbers, but what does this actually mean?

What does this model do that others can't already.

Valve just defeated Meta and Xbox in one go. nice.

Is there any way to edit programs in my own editor?

I can't work with these blocky fonts.

It must be a slow news day.

Some of them, but that's changing soon.

You bought the hunk of metal and glass, but you didn't buy the app store servers, the software has a legal agreement and you sign that.

If you don't agree, don't sign it.

These api models are for developers. Gemini is for consumers.

It's Apple's phone, they can do what they want.

A solution looking for a problem. We already have devices which can access AI with ease. I don't need your proprietary data farming apis in the OS level...

Just as any muscle gets weaker the less you use it, same goes for intelligence.

Apple has fallen off in every product category.

The naming convention is so messed up. o1, o3-mini (no o2, no o3???)

Very interesting. How can we get rich from this?

This is how it feels to own a desktop computer.

Page 20 has an employee with a HL3 shirt. It really is one of the most tragic tales in gaming history.

Half-Life Alyx only being available on VR was the final nail in the coffin. I will never play it.

So basically the model took user data and used it as training data in realtime? Big if true.

What do people even use iPads for?

What do you oppose with PWM, irregular subpixel displays? I'm curious why do you think these new improved displays are inferior to the LCDs of the past?

Mistral is really fantastic compared to other local LLMs. Even the 7B model is very usable on an older GPU.

Excited to see how things go from here in the open-source space.

Why do you need an app when you can watch it on the web browser?