Mazer23

Shameless plug: I’m workin on http://partforge.ai right now it’s mostly aimed at 3d printing but it can create laser cut sheet parts with dxf file exports and will support cnc routed parts with drill holes and socket diagrams built in in the near future. Board cut lists as a export is a little further out, but you can ask it for measurements of anything you build and it would probably be a pretty useful tool.

I feel like this is missing the for the trees a bit. I don’t think anyone who is saying “we don’t understand how llms work” means what this author claims. The underlying math of transformers isn’t the part we’re talking about.

What people actually mean is we don’t and often can’t understand an entire reasoning chain of a large llm and often can’t trace that back to the specific training data that caused that reasoning.

Anthropic itself had build all kinds of tools to try to understand llm internal “thoughts” https://www.anthropic.com/research/natural-language-autoenco...

I don’t think it’s unreasonable to say “no one understands how llms work.” What they mean is we don’t and can’t follow every one of its internal chains of thought, especially at the pace new models and techniques are being developed.

I spent quite a bit of time using openscad which Claude code to try to build a capstan (rope driven) joint for robotics. It worked well but had some issues with agentic stuff and speed. In the end I ended up taking a detour and building my own agentic cad modeling harness based on some of the same open source libraries that openscad uses. Now I’m trying to turn that into its own standalone product. You can check it out at http://partforge.ai

We tested this on our agentic CAD coding harness. This seems like a decent improvement vs 5.6. It's lower costs seems to be offset by more token use so those cancel themselves out, but the actual results for spacial reasoning and coding are an improvement.

https://www.partforge.ai/blog/2026-09-23-new-model-day

We tested this on our agentic CAD modeling harness. Opus 5.5 seems to be a pretty dramatic improvement in spacial reasoning and CAD coding. Our go-to is still Gemini Flash 3.8 for speed and cost reasons, but Opus 5.5 is a big update.

https://www.partforge.ai/blog/2026-09-23-new-model-day

Cool. It looks like they're trying to solve a lot of the same problems we are at Pixite. Unbound is our solution for this:

http://unboundformac.com/dropbox.html

I'm interested to see where they go with it and how they approach the problem differently.

This sounds a lot like the problem we're trying to crack with our photo app suite 'unbound':

Http://www.unboundformac.com/dropbox.html

All your photos on Dropbox and synced to all you iOS devices and macs.

It wasn't too bad, it took a few hours to get the basics up and running, then there was quite a bit of tweaking. The leap sdk is coming together but theres still a lot for developers to do. There are gestures but you still need to roll quite a bit yourself. I'm going to do another post much more in depth about the gestures and how to handle them in the next week or two.

Using print as an example of good high resolution design is interesting, but it misses one of the main reasons we have drop shadows and fake textures in non-print design.

Drop shadows and textures provide affordances and a visual language to communicate what a user can and should interact with on screen. They provide a visual hierarchy to the user so they can easily see what is content and what is interactive, etc.

Print has no interaction, so it's a bit naive to say we're going back to print aesthetics. We need a new way forward that provides both. Web apps have been experimenting with this for a while now. Some of these ideas work, some don't, but we can't just pretend print design has all the answers.

I'm pretty sure Airplay does exactly what you're talking about. If you play a video from YouTube or even streamed off your computer via airvideo, the apple tv connects to the server directly and the iPad or iPhone just acts as a remote. It doesn't use battery or network.

It's just seamless and you don't need to know how it works to use it.

Where did you study Interaction Design? I'm getting in a masters in it at CMU after getting an electrical engineering BS. The masters here is a great program.

My app is heavily effected as well. I have a Picasa client that auto-caches photos for users. The app manages the cache and clears it when needed, but certain albums can be manually set to cache by users so they're never deleted. I wouldn't want all these photos backed up and I don't think apple would allow it. Them going through and deleting the data themselves basically breaks the app for our users though.

The app: webalbumsapp.com

I wonder how the awesome sword and spade symbol on this website help him sell whatever he's selling.

The article you're referring to I think was written by Mark Weiser. It was called "The Computer for the 21st Century" unless you were thinking of a different one.

It's quite a vision for 1991:

http://www.ubicomp.com/want/papers/ubi-sciam-sep91.pdf

Sure, but ask any consumer what they think the best brand of tv is and it will probably fall between two or three brands that they can think of. I'm not saying a particular category can't be big enough to support a lot of brands. I'm saying in the consumers mind there are usually just 2 or three they think of.

The author seems to think that because there will be 20 or 50 viable androids they will somehow dominate. It seems to me that this is much more a branding/marketing problem for android.

Categories like this tend to support just two or three brands at the top of consumer mineshare. Think coke vs pepsi or BMW vs mercedes. The problem with android is it's not really a choice for a consumer. By picking android they really need to make a much more complex choice of which actual phone to get. The real choice the consumer is making is iPhone vs droid incredible or some other specific phone. In the end all the manufacturers under the android umbrella are competing with each other as well as the iPhone for mindshare.

I definitely agree that android will have more units out there than iPhone eventually, but I think the iPhone will easily be the brand with the much stronger mindshare in the category. Normal consumers will still see it as iPhone vs all those other phones. Apple will also have the biggest profit in the smart phone category as a result.

Remember, google is selling eyeballs to advertisers, not phones to consumers. They're playing a different game than apple.

Thanks for the feedback. I've definitely thought about the ability to add whole sites to the notes pane as a different take on tabs. This is a great idea but I'll need to work thought the implementation and figure out what the limits of webkit are. There are a lot of other things I want to add but this was an initial release to get a read on how people use the app and what kind of income stream I can expect.

Syncing is also a big deal. I'd love to use the simple note API as a back end if they ever open that up to me.