This makes sense - they are demonstrating the capability of their core product by doing so? They dont make browsers, c compilers, they sell ai + dev tools.

Capability of a product that makes non-working outputs at a premium?

I can hire an intern for that.

Will cost you a lot more ;)

Seems like a poor advertisement for their product if their shining example of utility is a broken compiler that doesn't function as the README indicates.

Impressive that it made a c compiler though? Or do we judge all programmers by their documentation now?

All it took was all the C compilers they could scrape into their training set.

It’s not impressive in the sense that it’s doing what it was designed to.

It just happens that it generated a C compiler that kind of worked.

Someone came by later and used more AI on it to make it closer to a production grade C compiler like gcc/clang.

Saying, “it made a C compiler,” is not specific enough.

Feels very negative! It costs 50% less than the air, in a time when everyone else’s prices are going up.

The single core performance smokes a lot of high end intel chips.

I don't think that was the intent. If anything, 90%+ of these features here feel nice-to-have, and I bet OP agrees (or is neutrally sharing the comp).

Feels negative because the positive of the cheaper price (it's half the price) is magically not listed

it reads like a 2026 version of CmdrTaco's famous iPod review, honestly

Better chip = better performance per watt = longer batteries for similar levels of performance, running cooler. Also it never hurts the smoothness of the interface.

It just ultimately makes it a nicer device to use.

But it's such an overkill. In any case, the question was a bit rhetorical, and highly subjective.

Interesting, a cool resource for an API endpoint for AIS data so aisstream.io. Seems quite solid. Any one any idea of a good resource for satellite AIS data - I feel like the EU probably funded it and I can’t find anything on capricious etc.

Super interesting to hear your experience, I agree that it is very dystopian. I have put up with it (with effect on my performance and somewhat my mental health), to be around my family more. Things like doing pick up and drop offs at school consistently has been wonderful.

Amazing story congrats on being a great coach - this kids will remember you and that experience.

My experience on tech as a parent (3 kids under 10), I find their time on iPads etc playing games, music and audio books to be good for them (they don’t get grumpy after it, and particularly playing Roblox with their friends online is great fun - real halo 3 vibes for me), watching shows they get quite difficult after if the have watched for extended periods (smaller the screen the worse it is), but if they get access to anything with a constant scroll / stream of things they go haywire. My son found YouTube on his nanas iPad and mainlined it for half an hour and then went crazy. My daughter lost it over browsing Amazon.

We are withholding social networks & scrolling video as long as humanly possible, but difficult when you don’t want them to miss out on anything, and there’s an element of controlled exposure…

Again great story, makes me want to sign up as a coach. Sorry for the tangent!

He would have understood the potential of LLMs straight away. If he would have delivered a successful product no one knows but 100% would have understood how LLMs could become another interface to computers.

I think his pitch would have been like garage band but for making apps “now anyone can make an app” feels very Jobsian.

There's this old gem about being able to ask Aristotle a question: https://www.youtube.com/watch?v=2YzLMPm3Jgw

I suspect Jobs would have held off for a few years just to see where the road is going before making a move. In that sense, Apples current lack of real dedicated to move into the LLM space has become a happy accident. They have avoided the hype cycle and the potential blow back once the more exuberant part of the AI craze fade away.

They have pulled an accidental Jobs move.

Why are you sure it was an accident? Couldn't the people who rose up the ranks around his orbit have learned the right lessons, and this having been very intentional?

I don’t think so, I think he would have thought Apple could deliver a better user experience (LLM Siri) and gone into it whole heartedly.

My perspective is that Apple doesn’t hold off to see what the market is doing. They hold off until technology can create a viable, usable product (iPhone was ahead of anything, Vision Pro less impactful but similarly advanced, iPod was smaller than any hdd based mp3 player - just some examples where they pushed the boundaries).

They 100% missed the transformer being a technology that could create a viable product, and I don’t think Steve Jobs would have missed that.

He for sure would loathe their current state though. They are a perfectionist's nightmare and detrimental to many brands. He wouldn't even allow Apple^Red for his friend's charity- there's no way he would let AI slop come from Apple products.

100% would hate to see the current state of Siri, and then even worse that others are doing what Apple can’t.

I agree that I don’t think he’d want slop, but I do think he would want Siri to be like ChatGPTs conversations, and he would have loved Claude Code.

It is insane that they have this little of a handle on their buildout. It makes the $600B feel even more empheral.

Do you think that Anthropic don’t include things like this in their harness / system prompts? I feel like this kind of prompts are uneccessary with Opus 4.5 onwards, obviously based on my own experience (I used to do this, on switching to opus I stopped and have implemented more complex problems, more successfully).

I am having the most success describing what I want as humanly as possible, describing outcomes clearly, making sure the plan is good and clearing context before implementing.

Maybe, but forcing code generation in a certain way could ruin hello worlds and simpler code generation.

Sometimes the user just wants something simple instead of enterprise grade.

I think that Claude can figure that out though - through the conversation Claude should figure out if this is an enterprise app, a hobby app or whatever, as it decides on how to code the thing.

I think this was more of a thing on older models. Since I started using Opus 4.5 I have not felt the need to do this.

Anthropic got rid of controlling the thinking budget by parsing your prompt - now it's a setting in /config.

They never parsed your prompt. The magic word reduces the probability that the token corresponding to the end of chain-of-thought will be emitted, which increases test-time compute.