verall

You can still backup to a large propane or natgas tank which many rural homes already have.

She is in it

agentic use which is mostly cached tokens really pumps the total token numbers without contributing to revenue as much.

These are basically all things that an engineering org can justify as necessary for some reason but can immediately disappear when leadership announces a change in strategy.

Not to take away from your point too much - this is a significant amount of work that shows their current stance towards maintaining the mac as a premium development platform. But I imagine they will start to split this - lower cost SKUs will get the iOS-ified OS while "Pro" SKUs will get (some of) the above and can still install homebrew, to prevent a full developer revolt.

Techies will figure out how to "jailbreak" the lower SKUs to unlock the full experience and the amount of people that follow through with that will round to 0%. Tech businesses will buy the >$3k Pros for their devs.

"Everyone" will be "happy".

This is interesting because it's the same logic either way -

"taking from people who work hard (the working class or billionaires) to give it to people who live off other peoples hard effort (billionaires or poor people)

In 2026 capitalism who is living off whom? Who is really toiling or creating utility?

Also to be clear, regressive taxes is ghoulish.

I see everyone saying this but I've been using 6-astra lately and afaict it's not much better

> You think you must have something wrong with you because you think nobody finds life and work boring?

> Struggling in life isn't an illness

Do you think it's possible that some people struggle 100x more? Should they not get help?

> Troi oi. Read what you wrote again.

I can get info from and even enjoy reading a biased take. It's not so hard to see where they are coming from. In the "semicon trash talk and rumors blogosphere", you take everything with a grain of salt, etc.

> Dylan is a grifting fraud. I wrote a too long document explaining a ton of examples and you're handwaving it away.

I've taken this to mean that it's your post so I went through it again. Some of the points are interesting but I don't think it's a very good case that he's a grifting fraud.

I enjoy the AI images done in their style though, it's pretty funny.

> AMD / NVidia put their best engineers on tuning, just for the benchmarks. It isn't about serving up inference fast.

NV has enough "best engineers" to have a couple people tuning for one of the most popular public tok/$ benchmarks without sweating. IDK about amd.

I dunno, it's a blog, so I'm not so worried about the motivation behind it besides how it biases their takes. I think being close to people that feed you information might be prerequisite to the kind of information he sends out.

I've seen the paid subscriber sections and it's nothing groundbreaking. I wouldn't/don't pay for it.

SBF used his altruism to cover up fraud. If the SA benchmarks were fraudulent, that would be a big deal. If he's just "in bed with the AI companies", like, that's a big part of the reason it's such a popular blog?

Also, I don't really have to think that Dylan is a good guy, and I certainly wasn't the only one that did not think SBF was a good guy. I get that it's really easy to call his implosion unsurprising after the fact, but it was truly unsurprising.

I really really disagree. Most of the prose is devoid of information. I find it hard to believe the author even read the whole thing one time.

I've been reading them since before all of the AI hype, and I've always thought they're pretty good. You a few spicy takes with the overview/opinions/benchmarks. Better than semiaccurate.

The article you link says not a lot of criticisms with very many words, and the AI prose gets much worse towards the end, seemingly when the author also gave up on reading it. I am disappointing in the plagiarism though, especially of Ryan Smith.

I am much more interested in what you think of the site though vs your own experiences running a GPU cloud. I've seen your comments on it for a long time, it's super interesting. So if you think their takes are mostly bunk I'd consider it way more than this hot aisle guy.

On large teams I think the "cherry pick" workflow (Gerrit style) beats the "pull request" workflow (GitHub/gitlab style). On smaller teams it's the other way around. I think it's somewhere around 10-20 people actively committing that the cherry pick workflow comes out ahead.

semianalysis is pretty good

Their parents?

Once upon a time the only thing I wanted for my birthday was 4 tf2 keys at $2.50 a piece.

It's not unreasonable.

Valve let's you sell your game on your own website and users can add the key to their steam for 0% skim. 30% rake for what is sold through their ridiculously popular store.

Apple should just charge devs for the services like fee per download if the app is free. I believe currently the minimum paid app is $0.99, so it seems reasonable to allow $.33 apps where the dev gets none and $0 if the dev wants to front the $0.33 per user.

I guess you could argue that even $0.99 apps are subsidized but I doubt it'd be accurate. They already charge to join the dev program and it's not that expensive to run an app store.