arjie

Very low quality content. Unsurprisingly pangram 100%.

In the end, it turns out for me that the classic explorer Tree View is the most intuitive. I know where things are as a sub-tree system. I get that hard links etc. complicate this and perhaps there are smart folders etc which do not fit this model well but I find I frequently want to “Go Up” in the hierarchy and this requires an uncomfortable shortcut on a Macbook (arrow key) or a menu button rather than a toolbar icon.

But these are complaints. The right answer is to see if I can build a file explorer for MacOS that behaves in the fashion I desire.

Absolutely agree. A modern agent is a sophisticated tool that can link multiple sources together to create a coherent piece of information hyper-specified for an audience. However, if humans have outlines that are expandable by LLMs it makes the most sense to do that as late as possible.

I don’t like reading AI writing either but I’m sure that’s a transitory period. Single prompt text expansions are unlikely to be useful because they’re late-bindable. You could give the original to me and I might be able to understand better.

But a series of steering prompts with various sources brought in is a different story. At that point it’s just a question of whether the agent can put together good information and their current inability to do so is unlikely to mean an inherent problem.

Some kind of UI affordance for this might help: with the agent emitting tags that allow for auto-folding or expansion in a way that allows both concise text and exposition when required by the reader. Mechanical sympathy, but for code executing on a human: good old human sympathy if you will

One could imagine a universe where the agent fills in citations and supportive points and so on, or makes a more conclusive argument but it seems you’d get better results leaving that to read time if it’s a one-shot. Steering prompts etc. with tool use to bring in other sources etc of course change this entirely. And regardless, I doubt we will read content like that directly ever again. Agents will act as per-person highly specialized adapter layers for information transmission.

The entire point is what runtime you’re running your code on. A computer with any modern stack requires a lot of text for you to communicate “spin a square around on its center” to it. A human requires only that short string because they have a faster natural language interpreter.

Text meant for a human can communicate “spin a square around its center” much better than any code that mimics it. In some sense, all programming is boilerplate expansion because computers have (until now) been unable to be programmed with anything approaching natural language.

Cities with the social order required for robotic street cleaning rarely require robotic street cleaning. One might as well suggest a novel heart surgery that works only on healthy patients.

Indeed, I was surprised to hear about “Lowell students” robbing people on Muni and “bicyclists” causing trouble in Mission Bay but when I observed some of this anti-social behaviour (when I saw them they were doing wheelies and swerving at people walking) I was rapidly able to discern why the statements were relatively vague as to the categorization.

I suppose it shall be the case for the short term at least that we must admit that “humans” really cause harm. Though perhaps that’s over specific and I should recognize that most crime is committed by eukaryotes.

I read an interesting take that it won’t. Because it won’t be interesting any more. It’s like how no one talks about AI IMO Gold anymore or Stockfish being better than all humans. This kind of mathematics goes back to being a curiosity of humans and machines move to the next frontier.

In a sense, the proof is a demonstrator not an end in itself. To mathematics enthusiasts it is significant. To the AI it is Tuesday.

Enjoyed that idea. Not sure how true but it was enjoyable.

> Tech industry circles nowadays are sometimes tinged with individualist selfish "take whatever you can" thinking, and anarchic or even underhanded methods. Including the thinking "an express no doesn't mean no, if i can find a way to take it from them anyway."

Is that particularly tech industry? Software and media piracy, ad block tools, etc. are all based on this principle.

Yeah, but most of these will fall. I used to use one or two of these to scrape my friends' posts because I stopped using Twitter and they'd inevitably all disappear in time. It's like chasing the latest TPB. In the end, I just logged in a headless browser and told an agent to fetch daily.

They are indeed "just next-token predictors". But rockets to orbit are also "just forward and upward goers". And Steph Curry is "just next three-pointer scorer" and so on. And as the article points out, chess engines are also "just next-move makers". A flawless oracular trading system is "just next trade maker". It turns out that "next n doer" is a true model that illuminates little.

Interesting question, and how much should a user pay Google to fix the vulnerability? I suppose the smallest unit of currency less than the amount of effort they'd have to put in to mitigate it. A fully market economy of bug fixing here is an interesting idea, certainly, but if I'm being honest I actually don't want to pay Google a thousand dollars to fix security issues. In the limit, what would happen is that I end up with the competitor browser Elgoog Emorhc which fixes security issues for free, and pays very little for them, which is the status quo.

In the world where security issues are paid for entirely at market rate, it would also be very important to not use browsers by poor groups because they would be unable to pay for security reports on the market and consequently the browsers would be less secure.

Interesting idea, for sure, but I don't think it lands in a place I want to go since I neither desire stochastic payments nor desire that all browsers should be from large corporations.

Seems to have taken it in good spirit:

> We shared the resulting proof with Kevin Buzzard, who said:

> > This extraordinary autoformalization achievement, which Anthropic researchers say only took 11 days, proves Fermat’s Last Theorem with no assumptions other than the axioms of mathematics. Along the way we see autoformalization of algebra, harmonic analysis, geometry and number theory, and we learn that AI autoformalization artefacts are now robust enough to be built upon; the proof is multi-layered.

> This plane cannot be used on any route longer than about ~200 miles. Which can be driven in 2 hours.

Huh, perhaps then it wouldn’t be suited to places where people drive at that speed. In the US I doubt people drive 200 miles routinely in 2 hours. It would be closer to 3.5 h I think going by the routine 6 h it takes SF to LA which is about 350 mi.

I think if you have very high speed road/rail transportation that dominates this across fixed routes. Road at 100 mph average definitely wins for flexibility.

Okay, well, that seems like a natural problem. I could understand if he went from one of the Gemini Flashes to the next (when they rebranded Flash to Flash Lite and came up with a new much more expensive Flash). Now that would be a mess.

The street conditions on Market St. are awful for bicycles though. I've biked that route very often and it was always a jangle machine. Now I have a trail bike with big tires and suspension and that's much nicer on it with the added bonus that I won't fit in the rails like that poor girl who died a couple of weeks ago on the scooter.

I think whoever the bike lobby is (SFBC?), they've succeeded in making things much more bicycle friendly than they used to be. Valencia, and even the new Folsom bike lanes are pretty cool. The Howard lanes are not so great but the fact that they exist is fantastic. The fact that the bike infrastructure is better in places where people didn't live (Treasure Island, the newer Mission Bay areas) points to the primary constraint here, I think, which is that most San Franciscans would prefer to keep things as they are. A city's ability to adapt to these things is mostly a function of its population age distribution and SF is mostly older people and few children and optimizes for that population primarily.

I remember I found an old video I took on a Sony action cam (useless device, it fogs up) and while all these places are still pretty bad, nearby Alemany does have green paint which is better than nothing. https://youtu.be/cYoJ4S7G5tc?t=98

Seems tough. The Ikea Saluhall is not far and since we live nearby we walk by[0] and it makes sense that that location lost many of its food places. There just isn't enough foot traffic. Alternative locations are much more popular. Stonestown is a much more popular mall for non-adults just because it's generally safer. For the SF Centre to become something like that it would require enforcement actions that I suspect would be unpopular in the city.

0: https://wiki.roshangeorge.dev/w/images/4/46/Julie%2C_Astra%2...

We were rescued from that dot org scam by the fact that ICANN is a California non profit. I wonder if the AG can lean on them again. This is an outrageous thing to do.

Considering a mail-in ballot gets rid of ballot secrecy anyway, I don't see why they don't just allow online ballots cast with login.gov or whatever.

Oh my god, I completely spaced dude. Weeks, not months. Weeks. Sorry.

Haha, entirely my fault but the default font greys out the text so I always just click through to the site instead of reading the Launch HN / Show HN copy. Regardless, I will take my well-deserved brickbats.

Familiar name. Same guy who made the masks during COVID-19?

Makes perfect sense. The classic way!

Yeah, they're good source. But the price for those GPUs is 5 figs even with the nvidia startup program nowadays. Also, I went back and looked. Most of my GPUs are actually from Central Computers who were great, but Exxact is real too. So "lots of" was inaccurate.

Also, the lead time I quoted was for individual 8x nodes.

You know what I'm curious about? Whether you have brand guidelines inside the company, a Claude skillset, or the blog post author makes the charts in line with the brand colours and so on.

I have quoted large nodes from this supplier and have lots of^W^W GPUs from them for personal use. Current lead time is more than 30 months.

They're a good provider but you have to be a big shot buying NVL72s before you're getting anything within your payback period.

This entire genre of pundit is just a person who has figured out that you can sell copium to the masses. Once upon a time this was a decentralized "this is how Ron Paul can win" thing on Reddit, but it was inevitable that slowly it would coalesce around these kinds of engagement bait people. Every subculture has its own such figureheads who DESTROYS opponents with FACTS and REASONING or whatever, but really it's just a reality TV show. The Alex Jones, Gary Marcus, and so on of the world are mostly just selling an entertainment product.

It's just a question of what the entertainment is. Some people feel good about being told "we were being hoodwinked; there are lizard people" and others feel good about being told "you'll be 100x healthier and good looking if you take these supplements" and others feel good about being told "these people are evil demons who are stealing your water" and others feel good about being told "these idiot rich people are going to lose their shirts" and so on and so forth. It's like how I like slice of life shows and hate horror movies and my wife likes horror movies.

In a sense, the misinformation gambit of LLMs did not come fully to fruition in the West (as much as it has in 3rd world WhatsApp forward land) because the locals were already a fertile ground of poor epistemic hygiene and well served by human providers of misinformation. Hacker News itself only has some hundred thousand commenters or so and even this requires a practice of aggressive information curation to prevent unrepentant misinformation repetition nodes from polluting one's belief set.

Where are you guys in SF? Would love to come see it. Mostly interested in seeing the speed of action (what multiple is the video replay at, etc.) and looking at center-of-mass etc. issues. I have a pretty straightforward use-case in a home. The Dreame we have in the living room is pretty good at avoiding most obstacles, but I usually have an LLM-based agent review the image automatically and either send out the vacuum or suggest moving things that would foul it up. Works great. Obviously substituting with a humanoid capable of reaching ground height things and picking up children's toys etc. would be great.

I'd want to bolt it with a mount to the wall when not in use because it needs to be climb-resistant for children during the daytime when I don't need it functional but there is a solid 8 hour 2200->0700 window when no one will be in the living room and speed is not an issue.

I was talking to my wife about this device and actually the wheeled form factor and (probably) no control software for balance is a feature for our use-case because when the Unitree G1 or whatever falls over the result is catastrophic. The balance control software attempts pure insanity and will likely kill a child nearby.

Name changes happen all the time and I would expect Google to match what the government sources use locally. The fact that the government is capricious is no reason for me to desire Google to become an alternative naming center.

Just in case you're not familiar with it, the reason there's a regulation against it is that a group of kids with the belief you're espousing lobbied for that regulation. They formed the Committee to End Pay Toilets in America. I suspect if there is a norm against this here in the US it was created around the time of that law.

https://en.wikipedia.org/wiki/Committee_to_End_Pay_Toilets_i...