ABS

you can pay at the parking meters directly, no need for a 3rd party app

for the on-device model, yes it runs on the Neural Engine (at the moment) so a newer chip means faster, cheaper local inference. For the server side path this Claude package is about your machine is irrelevant since it's a network call. The same API covers both, so "best machine for AI" only bites when the session is actually local.

But we can imagine that the balance of what's on-device vs what's remote will move continuously towards the former as time, improved HW and improved local models keep progressing

it's cloud, the doc is explicit that requests go straight to api.anthropic.com with Apple not in the way.

so Claude via FM dies offline while Apple's on-device SystemLanguageModel (the ~3B one) keeps working. It isn't a hybrid really: the framework just has both implement the same LanguageModelSession protocol so "local 3B" and "remote frontier model" become a one-argument swap.

IMHO what's worth internalising is that the two share an API but nothing else: the on-device path runs on Apple's Neural Engine and costs battery (you can watch ANE power ramp while it works) while the cloud path costs API credits/tokens and does zero local compute. Same code, opposite cost model.

also: Fable takes 2× the usage of Opus

something I haven't seen highlighted anywhere yet, while I find it very interesting, is the distributed inference across Macs (JACCL over Thunderbolt 5), an OpenAI-compatible mlx_lm.server, agentic-on-Mac.

Apple keeps MLX (bring-your-own-weights) separate from Foundation Models / Core AI.

looks to me like the docs don't give a feature-parity table, but they do draw the "role" lines once you read across them:

- Core ML narrows to classic, non-neural ML (its own docs now point you there for "decision trees or tabular feature engineering")

- Core AI takes neural nets and transformers (the new .aimodel format, the new profiler)

- MLX stays the separate bring-your-own-weights track (its WWDC sessions draw no line back to Core AI at all)

coreai-opt is the successor to coremltools on the optimization side.

among the many things I found that were not highlighted anywhere (that I could find) is the distributed inference across Macs (JACCL over Thunderbolt 5), an OpenAI-compatible mlx_lm.server, agentic-on-Mac.

Apple keeps MLX (bring-your-own-weights) separate from Foundation Models / Core AI.

lol, it took me 48 hours to do (and re-do, and re-do) this test + write it up and now that I convinced myself to stop changing bits and just publish it... Google's just announced the Gemma 4 QAT models :-D

It would not change the core of my article since the bottleneck remains the memory bandwidth on the old M1 16GB though

not obvious at all when every sentence uses "you" to indicate a general rule that applies to every Italian rather than "I" to indicate a personal experience

you are making the mistake of confusing your experience, which is of course legit but anecdata, with "how it works" in general.

I'm an almost 50 years old Italian so not a spring chicken but I definitely learnt Italian growing up, not a dialect, and not "from school".

I guess it's the difference between growing up in a city vs a village.

there is no lost love between me and Wall Street "bros" BUT it's not their job to run the companies, their job is to buy and sell shares based on results and potential future results.

When a company decides to go public and take their money it's still the company's responsibility to run their business and keep the people they took money from in exchange for shares happy.

as usual it's a matter of bubbles: most of those you know might hold such resentment, most of those I know are in fact in love with the possibilities and are trying their hardest to leverage the new tools.

My 64yo "non digital" graphic artist aunt holds a very high level of resentment to the digital ones, while most of her old friends and ex-colleagues who embraced digital way back when are stil active in the space and happy one way or another, she is not.

But she is happy she can now get near-real-time two-way translation to/from languages she doesn't speak and is also happy to bury her head in the sand when I point out that's thanks to the same tech that will have an impact on the people who do simultaneous translations as a job.

FWIW: Google says Android Open Source Project not being ‘discontinued’ amidst Pixel change impacting custom ROMs

https://9to5google.com/2025/06/12/android-open-source-projec...

Knowing the weight alone is not enough though: you can be 72kg and fat or 86kg and very lean but also very muscular.

E.g. I'm 178cm as well but when I was 71kg I was visibly "fat", or at least skinny fat to be charitable. I'm currently 67kg but very lean and somewhat muscular.

except you seem to think (like many others commenters in this thread) that "tech" means solely "silicon valley-type startups", which it does not.

I'll give you some examples that usually stun the average Italian hence they usually stun most other people as well. You think Italy and probably think fashion but in fact the Top 10 Italian exports are (first semi-random results):

  1) Machinery including computers: US$116 billion (17.2% of total exports)
  2) Pharmaceuticals: $55.5 billion (8.2%)
  3) Vehicles: $47 billion (7%)
  4) Electrical machinery, equipment: $45.8 billion (6.8%)
  5) Gems, precious metals: $25.7 billion (3.8%)
  6) Plastics, plastic articles: $24.3 billion (3.6%)
  7) Articles of iron or steel: $21.4 billion (3.2%)
  8) Mineral fuels including oil: $19.4 billion (2.9%)
  9) Optical, technical, medical apparatus: $17.4 billion (2.6%)
  10) Clothing, accessories (not knit or crochet): $16.4 billion (2.4%)

I wonder why US businesses and people import so much stuff from the EU if there isn't anything competitive coming out of it

no one would/should prevent companies from using also private channels like X, FB, Instagram, etc but enforcing a public channel that doesn't require private citizens to register, accept T&S and share their private data with 3rd party, unrelated private corporations to be informed of critical, public safety information would be helpful.

if only you had kept reading 2 more sentences after the one you quoted you'd know:

  "The license does not allow removing the original license and purport that the code was created by someone else. It looks as if large parts of the project were copied directly from Spegel without any mention of the original source"

except there are plenty of other restaurants serving maguro

assuming your point is correct about what almost the entire country thinks about ICE locking up tourists (and I don't think it is) it's irrelevant: ICE does it anyway and that's all I as a potential turist care about

if you need to eat and decide not to go to your usual restaurant for whatever reason... do you stop eating altogether?

start by learning the difference between Europe and the EU, then decide what you want to discuss.

With those two things together you can find the relevant official's office to call.

Or do you go straight to POTUS if you want to "speak to the USA"? (also note I wrote USA and not "America")

The Nazi salutes, the repeated commentary on individual European countries politics, his explicit support of far right parties in various European countries (at the time) upcoming elections. The related spread of misinformation at best and disinformation at worst via his x account regarding individuals in Europe he opposed, and so on and so forth.

I never wrote that Schengen "is not an EU programme", I pointed out succintly that the countries in Schengen and the countries in EU are not the same thing.

In fact there are also EU countries NOT in Schengen. And there are countries in Europe but not in the EU which are not in Schengen either.

I still have my Thinkpad from 2000 and it works flawlessly despite spending 15 years in a box in a slightly damp room (though the lid cover became sticky/tacky)

it's obviously from their own quotes but you can get most of the names in their various customers use cases, joint PRs and the likes (and those required the customers' direct approval )

"(or do not use only GCP)"

please read the text you quoted:

"Several non-EU countries are included in the area through special association agreements."

Schengen is different from EU.

More importantly, and according to your link, only Estonia rejected 50+% Indian applications, everyone else rejected less than 50%, with only 2 others anywhere near 50% (Malta and Slovenia).

So out of 29 countries in the Schengen area only 3 were anywhere near the 50% mark and all 3 are tiny countries as far as both area and populations are concerned (those 3 combined account for only 4 million people in total).

Also, just to take one of those 3, Estonia has an overall high rejection rate in comparison to all the others, and that started happening after the pandemic.

Details are important

"power of two" with "power" singular perhaps but "powers of two" with the word "powers" plural is really not common as a synonym of squared (excluding people who don't know what they mean at all in the first place)