sampton

Apple AI team launched Apple Intelligence before it was fully baked. That's a no-no by Apple standards.

Meta gave up on open weight path after DeepSeek.

Jupyter use is declining because coding agents got really good. Multiplayer mode is not going to save it.

Samsung has access to the same ASML tools but has much lower yield than TSMC. Chip making is hard.

MLX and MPS are 2 completely different teams within Apple. It's more like MPS team doesn't have control or visibility into PyTorch roadmap and can only contribute so much from their side.

The kid carrier looks dangerously top heavy.

You can train a cnn to find bounding boxes of text first. Then run VLM on each box.

Using building geometry to correct gnss signals is new.

There needs to be break through papers or hardware that can expand context size in exponential way. Or a new model that can address long term learning.

A lot investment is banking on agi. There’s no sign agi is going to happen this decade.

That's one way to lower the interest rate.

Those left coast elites.

More dense but less trust worthy. I don't think they would have pushed apple intelligence the way they did if there was a live demo.

I bet they rehearsed a dozen times and never failed as bad live. Got to give them props for keeping the live demos. Apple has neutered its demos so much it's now basically 2 hr long commercials.

If you want good prints and only care about the prints, get Bambu P1S. If you want to tinker with the machine and have a lot of free time on your hands then Prusa. although Prsusa starting to close up it's walls like Bambu.

Because Chinese government can tell their companies to adopt Chinese tech and they will do it. Short term pain for long term gain.

I don't that's true. Amazon sells infrastructure to other AI companies. If they jump into the model race they become a competitor not an infrastructure provider.

How is the material defined?

In the past few months I have used AI to read more open source projects than I ever had. Tackled projects in Rust that I was too intermediated to start. AI doesn't make you lazy.

1760000 BC: StoreTool 3. This is our greatest model yet. You are going to love it.

Short canva.

Servo + wasm.

People say that until they are asked to pay more and earn less.

It's not that easy fronting a job. I can't imagine one person going through all the interviews and juggle all the communication. Might as well just start your own agency with contractors from India or Philippines.

ChromeOS is genius in edging out Microsoft dominance in office software. Most interns only know how to use Google docs thanks to Chrome books.

Zuck swoops in and hire them 100mm a piece.

I like the idea but execution is underwhelming. For $500 put some wheels on that thing.

It's crazy we are hunting tuna to extinction yet here is perfectly good python meat going to landfill. Florida needs to build a marketing campaign to make wild python a delicacy.

I hope they can tweak the design for finder because the current beta looks bad.

My theory is the icon is more inline with liquid glass, where the translucent half goes into the background.