hahajk

There are many people who remember everything they have done, read, or experienced: those born with the hyperthymesia disorder. They'll remember every meal they've eaten, and the day they ate it, among other things.

There must be a limit to human memory, but we haven't found it yet.

Memories are highly compressed, anyways. So remembering all conversations and experiences probably requires fewer neurons than you'd expect.

And the draft FY27 NDAA bans all the rest of the Chinese labs for federal govt/contract usage.

No you're not. The article mentions downloading a model; talking about a local model. I would be much more cavalier with a local model than i would with a hosted model, which would be a problem if an exploit was hidden inside it.

This appears to affect users of Opencode Console, which appears to be an Openrouter-type thing?

https://dev.opencode.ai/console/

It's actually kinda hard to figure out what it is. But it does appear to be different that the CLI

From https://opencode.ai/v2/docs#connect :

> If you’d like easy access to all the best coding models you can try out OpenCode Console.

> you can't delete your credit card information from their system

In cases like this (for me it has been Verizon and Adobe), you can usually change your credit card on file. Nowadays I use Privacy merchant-locked cards for everything online but in the cases where I used a physical credit card, I would change it to a Privacy card and then pause that card or put a monthly cap on charges (after making a reasonable effort to cancel.)

I like to imagine walking into colonial Jamestown and trying to explain to them

"it'll be able to replace all physical labor with robotics fairly quickly, and we're left with no labor left for humans. We need to plan for that future or else billions will suffer from the results."

At least this brochure marks cooling in the "challenges" column rather than the advantages.

However, the only advantage listed in this announcement is "near-constant sunlight". It appears they are hoping for 8x solar efficiency for 1000x the cost of terrestrial construction.

If the actual motivation is that people don't want data centers on earth, that seems to be because the datacenters 1) drive up energy costs, 2) use a lot of water I guess, and 3) can make noise. If you're able to operate a datacenter satellite you apparently don't need local electricity, water, or human labor. As someone else said, set up on the desert or in the middle of Alaska somewhere 100s of miles from the nearest person. They can still use solar (just build like 8x the panels if you need to... still cheaper than space), they don't need water, don't need labor... I think most public opposition would disappear.

I've heard some say the real advantage is latency for spy satellites... This seems not a real use case. The kind of workloads that require this kind of datacenter usually can spare 1-2 seconds in round trip time. Millisecond-sensative reactions I would assume would want to be driven by onboard processing.

One true advantage that I haven't seen on any brochure is that the heat produced by the space datacenters could be released into space rather than raise the temperature of the earth. However, the satellite would have to be designed to radiate the heat away from earth. Hopefully they design the satellites to direct the radiated heat away from earth and don't just point the radiators straight back to earth...

> In its ideal form, this tax would capture and redistribute the annual rental value of land; that is, the recurring value of the land excluding the value of any buildings or other improvements on top of it.

> but you see the same basic patterns everywhere. Land in the city center is worth much, much more than outlying areas.

The calculation of the "unimproved value" always perplexed me.. it seems like you're not taxed for the things you build on your land, but instead for the things other people built around your land. After all, why would property in a city be valuable if not for all the high rises, subways, and office buildings?

This is not a problem with AI, this is a problem with a human accepting an answer without understanding why it is correct. The tutor could also read a book that states "A square is not a rectangle" (or vice versa) and repeating it back to the student as established fact.

Perhaps, but then someone would do the same but bill you for 7 hours, someone else would undercut them again, until the price reaches a lower equilibrium.

Another pro: they can easily support more than 4 connections

another con: If only one of the exits gets backed up, it clogs the traffic circle and causes every connecting road to back up.

You didn't counter with a non-humanoid design, your counter example is "many specialized robots" instead of one. What many want is a generalized robot that can do tasks the manufacturer never thought of, in an environment that already exists rather than adapting the kitchen to support the robot. The success of LLMs in generalized work over individual apps (the software equivalent of your counter example) has made people impatient for general robots.

> Oh, there's a driver update for the LG television. Makes sense, right? It's a fairly recent television, and the drivers will help us take advantage of all of the TV's features. Of course, I'd like Microsoft to download it.

At this point I assume the author was making intentionally bad choices out of morbid curiosity. I can't imagine anyone technical thinking drivers should be required to connect to a TV via HDMI. (or maybe I'm not technical enough)

> What do you think will happen if it isn't

Since such a thing has never happened, I doubt anyone, even scholars of history and anthropology, know.

Capitalism probably won't work in a world of ultra abundance though.

Doesn't the pixel tablet come with an unlocked bootloader?

Absolutely. If your account has access to the data, you can basically rewrite a website using a bookmarklet.

I wanted a friends feed in eBird (its social model is inscrutable) and ask Claude to make a bookmarklet that showed the most recent lists of my friends sorted chronologically. One shot and done.

Our internal corporate CRM is siloed data queries spread across overlapping views. I have a bookmarklet overlay my own interface on top and fetch data using the existing session token. (which is transparent/automatic - your browser just hangs on to the session cookies for the fetch calls)

A quick search for "labor supply" turned up this microeconomics lesson: https://youtu.be/002_HEC2l-4?is=Tu6dZfNOyK7DFpWB

Wages, like all prices, are determined by supply and demand.

If I understand correctly, NY is concerned about the power and water consumption. 1. How do they get power in space? (solar) Do the same on earth at 1/100 the cost (or build 50x the number of solar to account for atmospheric loss and you will still come out ahead) 2. How do they get water in space (they don't). Whatever they do in space to not need water, do the same on earth! 3. People don't like living near datacenters - put them in remote areas hundreds of miles away from people, after all space datacenters won't have employees . Again, at 1/100 or 1/1000 the cost.

anthropic's models would also be useless in a jammed environment though

Dispatch/Cowork is basically claude code in a container. The section "Why not run it in a container?" would address your question in the post. One example I run into is that Cowork won't download and fill out or read pdfs or other files due to container permissions. Vanilla claude code has no problem using curl and wget.

There is some evidence to support the idea. However, "learning" studies can be a little disappointing since they usually take place over a short period of time. This one had a week-long timescale:

https://www.nature.com/articles/s41598-025-97652-6

I think there's a bigger question that would be hard to answer: what does a consistent long term diet of videos and quick answers do to a person's mind/understanding of the world?

Also, a lot of the published research is a year delayed - most I've read recently are using ChatGPT 4 or 4o. When it comes to giving high quality feedback, I think there was an inflection point around Opus 4.6.

Sounds like they took a cue from Sesame Street:

> p8: "A child watching television under normal conditions is subject to frequent interruptions and distractions. The TV must vie for his attention. In order to simulate this condition, we decided to program distractions into the laboratory situation... Slides could be used to fill the slide tray and they could be projected automatically, at regular intervals, onto a screen similar to that of the television set. The carousel projector allows the viewer to choose three exposure times. The 7.45 second interval proved most satisfactory with the preschool children.

THE FIRST YEAR OF SESAME STREET: THE FORMATIVE RESEARCH https://files.eric.ed.gov/fulltext/ED047822.pdf

You're talking about an industry that gives you a generic "check engine" light and allows dealerships to charge you $100 just to read you the error message.

I don't have any faith that they have the wherewithal to make any common sense design choices about error states.

Related: Science is a strong-link problem

https://news.ycombinator.com/item?id=35712694

Things like lottery tickets are "strong link" problems (the system's value equals the strongest link's value); food safety is a "weak link" problem (the safety of a meal equals the least safe bite of food).

It appears that the author is advocating for "weak link" standards for CSS (it matters more what the vast majority see, knocking it out of the park once or twice won't make up for the failures)

I've never used Gemini computer use, but I assume it's the same:

Claude computer use takes control of your whole computer inputs (mouse and keyboard) plus screenshots. You just log in, tell Claude you're logged in, and let it get to work. It'll use the browser you're logged in with.

The chrome extension is a little better because it only takes control of its own chrome tabs (again: you just log in.)

Someone has to archive the data, right? Perhaps the complexity has been pushed to the AppViews?

If I read a post from a month ago, how will my client know how many likes it has or what the replies were? Without reverse links, someone has to look up in a database of all AT actions for likes/replies pointing to that post.

Contrast with the inbox/outbox forward/reverse linking model of activity pub.

I've been setting up some federated servers (Matrix, activitypub) and I ran into .well_known/ paths in many of them. Webfinger resolver for activitypub and a more custom matrix server-to-server federation endpoint.

> Europe already has an ATproto social network - Eurosky

Why is this a different network? Are Eurosky relays not indexing anything outside Eurosky?

It's not necessarily that simple. Both I and my wife read, almost every week I read my daughter passages I've come across in my books, we have bookshelves in every room. We read to our younger children and they enjoy being read to.

I told my daughter that when I was her age I liked to read Animorphs, and girls were reading Babysitter's Club. She brought home these books from the school library and... they were graphic novels.

Apparently the school library is stocked with comic books and the kids can just read those instead of real books. And comic books don't have descriptions of scenes, they have almost no internal monologue or exposition, no symbolism or (literary) imagery, they really can't teach reading comprehension.

> interest you pay is insuring the bank against bad debtors

I thought that was the govt:

> Loans securitized by government-sponsored enterprises (GSEs) like Fannie Mae and Freddie Mac continue to dominate the market, comprising around 52 percent of all balances at roughly $6.5 trillion. Government-backed loans, such as those insured by the Federal Housing Administration (FHA) or Department of Veterans Affairs (VA), account for 19 percent or $2.5 trillion. FHA loans are designed for first-time and lower-income buyers and make up 12 percent of balances, while VA loans that are available to U.S. military veterans comprise 8 percent.

https://libertystreeteconomics.newyorkfed.org/2025/08/a-chec...