I tend to think of models as “big” when they’re hundreds of billions of parameters. GLM 5.x is “big”. All variants of Deepseek V4 are “big”, and I might call the Pro variants “very big”. Kimi K3 is “very big”.

I do not own the hardware to run those locally.

This matches my experience when I’ve used midsize models connected to read-only tools.

It does not match my experience using a big model in an environment where it can generate and execute code when given instructions to base its answer on that. That pattern almost always results in a factual answer, and often, but not always a good one. The downside is you need to set up such an environment or pay a provider for one, and it’s slow. If you have something else to do while you wait, it may improve your efficiency. If you don’t, you might as well do the experiment yourself.

The opinion I’m starting to settle on is that expecting people to read LLM output is rude.

That means at the point something is released for public consumption, the documentation should be human-written. At the point an open source project solicits collaborators, the code should at least be human-reviewed and probably human-refactored. I’ve read enough LLM-generated code to know it isn’t very good at picking abstractions.

There are, of course business school types in industry who want to avoid having humans do any of that work because it’s expensive. This leads to the same kind of problems any other decision to reduce quality to save money does.

I don’t know what the solution is.

I do. Eliminate the border search exception for the contents of digital devices. It’s already trivial for any serious criminal to transfer illegal data over the internet instead of carrying it on physical devices, so such searches are usually a pretext to look for something else.

Does Reddit really have any less engagement than Twitter or Instagram?

Yes, and no, respectively, according to Statista.

I think the more algorithmic platforms are trying to maximize performance for users who shut off their critical thinking and consume whatever they get, especially ads. Something more forum-like doesn’t fit that model.

Most do. Source: I usually use passwords that long.

Maybe 1 in 10 don’t, and I find that annoying since 32 characters containing only lowercase real words and spaces has several times as much entropy as 8 characters of random mixed case letters, numbers, and symbols.

If your security relies on a password, you will always be anxious that somebody, somewhere, has somehow figured out what it is and already has access to your bank account.

No I won’t. All my passwords are something like I0!2b6KJ!eA04WiJ&&MFToCQ7CN*4mvb and each is used for only one thing because I’ve been using a password manager for over a decade.

Any of the following are likely to work:

• Web search for “[model name] linux” and see if there are any results describing actual experience. A handful mentioning using it with none mentioning compatibility issues is usually good to go.
• Check forums for existing threads. Good options include https://forum.thinkpads.com/ and r/thinkpad. Less populous possibilities are !linux@lemmy.ml and !linux@lemmy.world.
• Make your own post on any of those places and ask about the model and distro you’re interested in.
• Use an LLM-based tool, but check its sources to make sure they say what the LLM claims they say and the author shows signs of knowing what they’re talking about. LLMs are bullshit machines.

I guess then I would add to my advice that people should do a quick check to see whether others are having good results with a given model before purchase. I thought we were past that with Thinkpads.

I’m surprised to read that; I would have expected substandard build quality, not problems with Linux.

Ctrl should be on the key that says “CapsLock”, and for me, it is.

You can swap Ctrl and Fn in the BIOS.

My newer model can kernel panic as well as any model!

I had a T470 previously, and the mechanical build quality and materials of the P14s III are superior.

There’s a reason Linux on a Thinkpad is a meme.

You haven’t said much concrete about your requirements, use case, or budget, so I don’t have a recommendation much beyond whatever Thinkpad meets your requirements. The “serious” models in the T, P, and X series tend to be preferable to the others.

I love mine!

I’m used to people trusting friends and family members who they know are tech experts. That’s a really weird reaction that seems disrespectful toward your knowledge and judgment; you weren’t wrong to be mad.

Similar situation, except involving oldies (70+).

Every 70+ family member I have asked to install and use Signal has done so without any assistance or instructions beyond the name of the app to install. It’s one of the easiest tasks in modern information technology.

It always surprises me when anyone resists. It’s 45 seconds worth of effort to install and activate it, and anybody who knows me knows I’m not going to recommend software that will harm them.

When governments tried to regulate digital stuff 30 years ago, we got anti-circumvention laws that let corporations put digital locks on everything that are illegal to break. It would be hard to convince me governments regulating social media 20 years ago would have been anything short of a disaster.

Social media run by corporations based on algorithms designed to maximize engagement was a mistake. It may, however have been inevitable.

By that I mean I’m not sure a clever person with a time machine could go back 25 years and prevent it from having roughly the popularity and influence it does now even if someone with a quarter century of future knowledge tried to outcompete it with something that isn’t centralized or optimized for engagement. Those traits were evolutionary advantages over that time period.

I was tempted, but nobody learns anything from that.

I have the impression the author of the article has dabbled with Lisp but hasn’t actually set up a reasonable development environment and written something non-trivial. That’s not a great position from which to write critiques of a language.

To add to the point about homoiconicity, both the article and my toplevel comment mention thread macros. The article talks about them being equivalent (syntactically at least) to method chaining, which I agree with.

The thing is, to get method chaining, the language has to have an object system, and the functionality you want to use has to be written in an object-oriented style. To get a thread macro, you need ten lines of code (aside from the docstring and metadata).