Any accuracy measurements?

I majorly compared it to the native Explorer agents (for example in claude code). So far it has won against the explorer agents in 98 of 100 cases. I am already in the works to create a bigger benchmark, but did not have so much time for it. But you are welcome to test it out :)

I'm sticking with Sonoma as long as they still provide security updates, then i go +1, hoping to skip the whole liquid glass phase.

It was an anecdote or a story shared on HN eariler this year.

There's a (TED talk?) reel been doing the rounds the last couple of days. One of those "interest dies down, someone finds it again, shares it, there's a new bloom of viral interest, repeat until heat death" things.

If you built it FOR someone(a business), you are already on the wrong path. You should have built it WITH someone(a business) and you'd already have feedback and real world usage. Maybe next time around start with a real problem (confirmed by a business) rather than how you want the world to work.

The critique of metadata being hard is fair, the claim that sealed sender is “totally useless” is not. It’s a small, incremental hardening step in a very messy design space, not a magic invisibility cloak, and judging it as the latter sets the bar unrealistically high for anything that still wants to be a drop-in WhatsApp replacement.

It's useless in the sense that it makes an anonymity promise to users that it cannot fulfill.

Just because it's not perfect, does not mean it useless.

A central signal message service receives millions of messages, I've seen claims of 40M active users a month. If each user sends 25 messages a day, that's 12,000 ish a second.

Drawing conclusions about who is talking to each other out of a 12,000 message per second stream is far from trivial since both signal users are just sending and receiving encrypted packets to a central service. Much depends on how much you believe about how signal handles things on the server size.

Not sure federation or pure p2p would improve things, especially since some fraction of the service could be malicious.

But it doesn't make the claims OP says are broken. Op makes several logical leaps and because each leap is a reasonable leap, he assumes it must be THE leap. Which isnt true and it's simple to come up with counterfactuals, and it's a common pitfall in analysis (he's confirming his bias)

attention is all they want

Attention is all they need.

Can't we just give them a hug instead if they're that lonely?

It's about the new trend of shallow DoF in new movies vs old ones.

I think this is close, and the video touches on that as a characteristic that’s contributing to this, but there’s a motivation left unaddressed by the video that needs to be called out:

Reducing depth of field reduces the render resolution, which reduces the costs of digital processing and generation.

The simplest way to demonstrate this on a desktop computer is with the photography mode in games like Minecraft or Satisfactory or Elite Dangerous or No Man’s Sky, where the user can modify the Render Distance and Depth of Field at will. Load up the game viewing some planetary scene and enable the fps counter, then start changing the render distance; the closer you set it, the faster each frame will be generated. But the background will look defective and empty, so add depth of field, and now it doesn’t look so cheap — and when you take the photo, depending on the game, it may override your realtime render distance because it can take five seconds (!) rather than 1/60th of a second to generate that frame at 20 megapixels.

I think that the shift towards low depth of field in movies is, in part, a reflection of cost pressures, especially in 99.9% CG movies like Quantumania. And I think this is where Avatar beats out the competition for pure CG worlds in this video, because it renders at full resolution. It must have cost significantly more to produce than Quantumania (yep, $250m > $180m). I wonder how much of that difference was due to rendering the entire movie with a cheapness DoF blur. If nothing else, shadow rendering is so much of the difficulty of CG, that it could plausibly alone be the reason.

(I think that low depth of field is also currently popular because mobile phones lack it, and so producers are consciously or unconsciously selecting for an experience that is distinct from what they might film on their own. Depth of field is a very cheap form of escape.)

I think that is the wrong lesson to take away from the video. As the video emphasizes, DoF is a tool that can be used to achieve an intended effect in story telling.

Main thrust of the video is that these days these tools are predominantly being used for convenience of post-production and cost cutting at the expense of immersion and story telling.

I use a regular 3kg 17” macbook pro from ~2007. Beautiful keyboard, good enough resolution, wifi off(not much use on the internet anyway). Still modern ux and good trackpad.

The original iPhone.

100-700$. more when actively building, less when maintaining.

That's crazy. What kind of stuff are you building