apgwoz

> Anthropic claiming that its total revenue since January 2025 as $5 billion contradict that its expected run-rate revenue for the year 2026 is $19 billion?

Isn’t the “exceeding $5BN” comment a lifetime revenue? … on $30BN (edit: previously said spent) raised (or something ridiculous.)

A lot of the commentary on the frontier model companies is based on how much money they’ve spent to the relatively small amount they’ve made in return, and the skepticism, especially given almost continuous reporting, that deploying AI in a variety of situations doesn’t seem to yield favorable business outcomes. OpenAI shifting to enterprise / coding type stuff this week seems, also, potentially informative. Is Gen AI actually useful for anything but code? Signs keep pointing to no… and even then, we’re in the early stages of figuring out how to build without destroying everything… something Amazon just recognized as possible with their recent shopping outage.

> There’s probably more I built that I have already forgotten about.

This is a big gripe of mine at the moment. I rarely have any confidence that I know how the thing works, or what additional things it does / does not do but which I expect.

Recent example: all API endpoints should require a bearer token. Imagine my surprise when half of them didn’t enforce this effectively, 3 days later. A bearer token would work, but also providing no bearer token would also work. Over the course of time, tests were removed / things were modified to get to the goal and say “done, boss!”

I’ll note that for this project, “don’t look at the source code” was a requirement. Things have been corrected before release, but the amount of potential foot guns is so damn high.

> with their demands that the government already knows exactly where they live, where they hang out…

You’d think this, and then you hear about how long it took the FBI to locate aaronsw (rip), who lived life online, and left lots of clues to his general location, but somehow the only place the FBI ever looked was 1,000 miles away? I guess you could say that was 15 years ago, but we had domestic spy programs 15 years ago, too.

The irony of an ex-Google engineer coining Hyrum’s Law (https://www.hyrumslaw.com/)

So, if I use my SIM card 16 hours a day, 7 days a week, Ill get banned? Doesn’t that seem absurd? The SIM card is enforcing one voice call at a time. If the apartment building has to wait in line to use it, what’s the difference?

If you deployed it in a way that did multiplexing such that multiple users could use it at once, then sure—-Business time. But otherwise…

Huh? I know how it works. My point is different than the representation.

No. The quoted text says nothing about the value semantics, only the representation…

There have always been attempts at caching non-string values in Tcl for performance gains, but the semantics are the same. The value’s types are whatever the operator requires and an error when not compatible. If, internally, $x is an integer, and you pass it to a string function, well, $x gets temporarily turned into a string. Dynamic with “weak” types.

Easily the most quoted part of the film, aside from “Hack the planet!!!!” … but also an amazing prediction! All the devices in our pockets are RISC machines. That did change everything.

Can text be put on a screen? And can you take a picture of it when it is? Well, you might have a screenshot of text.

Yes. Except they are attempting to make the claim that protesters are interfering with federal operations, which is a crime. Therefore, they can try to make the claim that they are only investigating potential involvement in a crime, punish you, and file it under a violation of the terms of service for precheck and global entry. IANAL, etc, but this seems to be the strategy.

Any good and honest tansu experience reports out there? Would be nice to understand how “bleeding edge” this actually is, in practice. The idea of a kafka compatible, but trivial to run, system like this is very intriguing!

Perhaps I am missing something, but this seems to require a Lemon (LLM)? Is the idea that the Lemon is used to help build an index AOT that can be queried locally, after?

I want to figure out how to build advanced tools, potentially by leveraging Lemons to iterate quickly, that allow us all to rely _less_ on Lemons, but still get 10,20,30x efficiency gains when building software, without needing to battle the ethics of it all.

Not to a professional, no. But this isn’t iMovie vs Final Cut. Aperture was only slightly above where iPhoto was going in capability. They should have raised the iPhoto/Photos bar a lot more to get back to where Aperture was, though.

I do like the convenience of iCloud, but totally agree that having them safe elsewhere is necessary. I’ve been pretty bad about keeping solid, non-iCloud backups of my photos. I definitely need to be more proactive about it.

I think Apple killed Aperture primarily because it was confusing to have iPhoto and Aperture with largely overlapping workflows. Aperture had the loupe view, and side by side comparison stuff, saved color grading tools (I think?), sure, but it wasn’t differentiated enough to justify a Pro designation. I think it makes more sense for Photomator features to be absorbed into Photos… and maybe Photos gets some new Pixelmator integrations if you have it, for quick touch ups / enhancement type things.

On the other hand, Final Cut / iMovie will exist side by side because it’s truly a basic vs Pro situation.

Not a product manager at Apple, of course, but this is what logically seems to make sense.

I’ve been waiting to see what happens with Photomator, and the fact that it’s not being included in anyway here makes me think it might not survive? Either that, or it’s gonna be heavily integrated into Photos…

Downvoted for the reason that I’ve not seen any credible evidence of massive election fraud, in the USA, since the suggestion of it in 2020. If you can point me to something credible, that proves even 1,000s of votes in a single district were impacted, (thereby stealing an election via mail fraud) I’ll remove my downvote.

I think it might be _unethical_ to not spread the joy of playing Doom for the first time? Though, I’m not entirely sure there’s been enough research done about the effects of violent video games in rat gamer populations.

I was very happy to see this. I’m fairly against live animal testing, but giving rats the joy of playing Doom??? I think I _may_ have to be OK with this.

Not to mention HashiCorp bled talent before the acquisition was even announced (BUSL started it) and it didn’t really stop as far as I’m aware.

Heatmap based on coordinates of the start of the boing!

Wait. Is Jelly Car basically a rethinking of this? I never managed to have the elasto games, but looking at the trailer, there’s a lot of similarities.

The onus is on you here… but, I think I know where you’re going with this. In terms of number of email addresses people have and use, vs number of usernames people have and use, you might be right that some people have 1 or 2 email addresses and many usernames.

Email masking has become easier to use, and many people use `+addressing` to uniquely tie their email to the service for spam prevention / tracking, which would make stuffing harder.

In these cases, email would be much more unique and a better protection against stuffing. HOWEVER, it’s not obvious how Email verification protocol would work for these types of things.

I think that the "tools" movement is probably the most interesting aspect of what's happening in the AI space. Why? Because we don't generally reuse the "jigs" we make as programmers, and the tool movement is forcing us to codify processes into reusable tools. My only hope is that we converge on a set of tools and processes that increase our productivity but don't require a burning a forrest to do so. Post AI still has agents, but it's automatically running small transformations based on pattern recognition of compiler output in a test, transform, compile, test ... loop.... or something.

You’re not wrong to bring up the social level, but you also have to combat the technological level as well. The systems that already exist don’t work not just because of social reasons. Mastodon/ActivityPub was gaining decent traction, despite its usability problems … and then the more usable, better funded BlueSky stepped in.

There are lots of options for running your own Mastodon/ ActivityPub implementation, etc—it’s not easy, though. This is purely a technical problem.

Of course, I am not really talking about Chirp clones here, but this is a pretty good analogue for comparison.

And here I am doing this as a joke (https://sigusr2.net/serving-traffic-directly-from-sqlite.htm...)…

> people just don't use it.

… which means it’s a problem.

That’s exactly the point. Right now, “independent media” (across all political spectrums, mind you) is trying to rise up, but they’re doing so on platforms that _could_ easily turn them off. There’s a ton of multi-million subscriber Left leaning, independent media channels on YouTube that could easily become irrelevant by a simple “change to the algorithm” that squashes them.

Patreon, Substack, etc are no different. We’ve seen people be silenced on X. We’ve also seen payment processors disallow payments to organizations and whatnot.

This is all a problem.

It’s less about being susceptible to influence by the rich. Nothing can truly stop that. It’s more about ensuring that platforms can’t silence dissent.

You can dissent in an echo chamber but it doesn’t do any good.