trvz

As a European: I don’t care where an open weight model comes from.

I don't disagree with the title of this weblog post, but its premise is wrong: the existence of AI has an impact on the total number of tech companies, but not on the number of companies that make it real big (the AirBnBs, Stripes and Dropboxes the author is looking for).

For most people employed as programmers, inheriting code from previous people has been a thing.

Also relevant: including libraries written by others in your own software.

As always it comes down to finding a healthy balance.

You misunderstood what I was referring to by “hacker” there.

First, smaller models are fun for hackers: you can run them locally, or run them faster.

Second, when cloud models become unavailable or otherwise deteriorate, these will be all you have. May as well prepare.

All M3 variants use LPDDR5.

That M3 had an older type of RAM. Apple hopefully secured sufficient supply of the newer variant for the M5 Ultra.

It’s implied that they do, but don’t have the balls to tell you they do.

"matches" could've also been written as "this year's product falls short of the competitor's last year one".

And multithreaded is 10 cores vs 6 cores.

I appreciate Apple being kept on their feet, but they're not dethroned yet.

Strix Halo is 256 GB/s.

You used the wrong units there. Should be Mbps and Gbps.

It’s not even greed or corruption – it’s mental illness.

I think the appropriate way in this case is to cover it up with black tape.

> How this was written: drafted with an AI assistant from my own notes, then edited by me.

Chill, dude. The AI datacenters generate enough heat.

Literally the same security risk.

“High parameter […] models” was mentioned above. It’s still not a good take however, as laws of civilizations getting more convoluted over time is the general behaviour, not specific to LLM-producing countries.

Agentic behaviour isn't just coding. In fact:

> We recommend using it for agentic workloads, tool use, data extraction, RAG, and long-context workflows. It is not recommended for agentic coding and knowledge-heavy tasks.

No. No.

They're not. When things go wrong it's better to compromise someone else's VM host than your own computer. It's only a matter of time now until AI will find novel ways to break out of virtualisation.

A VM hosted by someone else. Somethig on your personal notebook or the proxmox server in your garage is still too risky.

It’s a common word amongst people with even a bit of erudition. In truth, using it for any kind of software is overusing it.

The main argument you give for it not making sense can easily be changed in Settings. The energy and network considerations have no impact on running a virtual machine in UTM – iPhones will give everything they can as long as thermals are fine.

Yeah, but it’s also plain greed, and it’d be hard to tell the ratio.

Having to pay for stuff you use is not a dark pattern.

macOS covers 90% of people who would ever pay him.

It's implied. Truly smart people manage to become affluent.

Soldered RAM being bad is an opinion, not a fact.

I’ve seen way too many RAM slots go bad.

Smart people always get the highest configuration in computers, unless they know they’ll never need some of the resources and can tolerate a lower option.

Especially when the movie is over forty years old.

Most of the big VPN companies known from advertisments all over the internet are probably honeypots of the usual countries.

It would've been surprising before LLMs. Now that you could run all of your communication through AI, or at least have it check it, it's outright negligent.