So, in a way, we're more likely to see the Dune version of "machines taking over the world/universe" than say, Terminator's even though many joke about ChatGPT/Claude being Skynet.
It made for a great marketing stunt, it is absolutely useless for a production car. And you can tell the designers realize it but were at the whims of Elon in their interview.
We were promised an awd, 620 mile range monster almost a decade ago, and then again 6 years ago. We were delivered a “maybe 310 mile” range, rwd doorstop.
I’m not sure who this car is for, but it’s definitely not car guys. It is the cybertruck of sports cars, and that’s not a good thing…
However, the COPV-powered reaction thrusters could shred anyone on the wrong side of the vehicle. Even in the non-invite-only 0 to 60 in .6s version, that is also a problem, correct? Meanwhile:
> Tesla 'Cannot Guarantee' Your Roadster Pre-Order Will Ever Be Delivered
https://insideevs.com/news/812247/tesla-roadster-preorder-de...
- Ordinary term used more than once in the same comment
- “thrust” appears 2 times
- “rocket” appears 2 times
No, both versions use cold gas thrusters. There are no combustion rocket engines involved here.
The difference between the “thrust” and “rocket” versions is the number of thrusters: “rocket” has additional thrusters that enable it to fly. The “thrust” version just has them at the back for acceleration.
- Quoted phrase in the opening sentence
- Opening quote: “Just a good ol' boys, Never meanin' no harm”
[0] https://www.youtube.com/live/qt5p-_tE8d0?si=V5csPghK02wwd0or...
The problem with the demo is that it was vapourware. Period, full stop. It is not the car they are selling to us, and they haven't even begun making them. I am personally not a fan of pure concept products that exist only to polish the company’s crown.
Steve Jobs said that "Real artists ship," because the act of shipping is the act of making the incredibly difficult and important decisions that manage the enormous number of constraints that go into the actual product.
If people like concept cars that don't have to actually work or meet a specific price point, fine. It is entertainment, nothing wrong with that. But I am not and will not be impressed until, like Apple, the thing they show on stage is so far along that you can be expected to buy one within weeks, not years.
I am not telling anyone how to judge this. But by my very old-school standards, vapourware sucks.
This is explained in the deep dive video with Lars - https://www.youtube.com/watch?v=22rvpUecLts
“Play” “fun”, curiosity” and “challenge for challenges sake” are adaptive strategies for learning unexpectedly useful things.
And inevitably two groups will clash, one pointing out how much this deviated from the original, often decade long promises vs another pointing at the identical naming, how it's still a product they like and of course the all time winner "I don't see you do X"...
For what it's worth, I'd file this as a RCS demo for SpaceXGrokSI (or whatever they call themselves this week) above all else and if I were a betting man, might be a prelude to fusing them into SpaceXSuperXTeslaXIntelligenceX...
There are lots of ways for sleight of hand. Do you remember the time that guy made the Statue of Liberty disappear?
- Quoted phrase in the opening sentence
- Opening quote: “actually flies”
- Quoted phrase in the opening sentence
- Opening quote: “normal”
- Quoted phrase in the opening sentence
- Opening quote: “multi-mode suction system.”
- Performance: road ready, seats two, active aero and a “multi-mode suction system.” 0-60 mph in under 1.6 seconds. Available by reservation.
- Thrust: race ready, seats two, adds rear-mounted cold-gas thrusters. 0-60 mph in under 0.9 seconds. Limited release.
- Rocket: seats one, adds thrusters around the perimeter of the car for upward and rotational movement. 0-60 mph in under 0.59 seconds and a top speed of over 230 mph. Invite only.
Per Electrek: https://electrek.co/2026/10/15/tesla-unveils-cyberfied-new-r...
It's an incredible engineering achievement to build in massive fans and rocket engines into a high performance car. It's truly wild.
It's being sold in two modes. One with cold gas (no combustion) seemingly just for extra forwards acceleration (no flying) and one with real combustion rocket engines, the latter was in the video and is "invite only".
Wild that they devoted so much effort to an economically completely infeasible model. I wonder if the combustion version will even break 3 figures sales.
Even just the cold gas variant seems like a nightmare for production, there's a high pressure COPV in that thing and that's gonna make crashes much more.. exciting, and Tesla doesn't have a great record anyways. But very few of these are ever getting sold anyways.
I'm a Home Assistant user because I don't like my devices reaching out to random cloud services more than they need to. I suspect quite a number of fellow HA users feel similarly.
Although your intent was probably good, this feature certainly falls into the "more than they need to" bucket and feels hella icky. And yes, I'm very familiar with HIBP and cryptography in general. This simply isn't an opt-out feature I expect in a privacy positioned home automation product.
- Should a password manager send even one byte of my data to any service, regardless of whether it is for a good intention and mathematically proven to be "safe" without opt in? I'd say no.
- So would any other product out there, be it a smart home manager or a toaster, send any of of my data, hashed or not to any 3rd party without opt in? even if it's just the byte "1" to a page called "donothing.com"? I say hell no.
- Should open source developers who are not Linus Torvalds be a bit nicer on a public issue?
- And should open source users should be less heated because "let us know if you plan to fix this or we will uninstall and fork it" deserves the "go ahead, this is open source, threats don't really help here"
We all can sometimes write a reply at the heat of the moment that we later regret. Frenk definitely realized it after the fact and I respect him owning it, apologizing, and keeping the old messages. He is right it's open source, and he doesn't owe anyone anything, but still, I believe a phoning home of any kind is violating The Principle of Least Astonishment.
You might have a heart of gold, and just be having an off day due to reasons, so I hope people cut you some slack, but if I were you I'd update this feature to be opt-in and conspicuous, immediately
- Quoted phrase in the opening sentence
- Opening quote: “you”
- Quoted phrase in the opening sentence
- Opening quote: “the exact set of people who read Hacker News”
At the risk of stating the obvious, this is a forum where people can post arbitrary links to other sites, some of which might have audiences that might not be "the exact set of people who read Hacker News". I'm not sure you'd be able to make a compelling argument that alexwang.ai is aimed at software developers and no one else.
They're not even in the same ballpark - the PS5 is more than an order of magnitude better in GPU performance.
Despite how much it does matter, RAM is not the only thing you need for running LLMs - that N150 is not going to run anything but really tiny models at any sort of reasonable speed - certainly nothing that approaches using all of its ram. The GPU in the PS5 is ~25x faster. And if you're running games the RAM doesn't really affect performance at all once you have "enough" for a given game.
You're right that you probably could build something quite a bit faster than a PS5 for the same price though - if you can build/find everything but the GPU for ~250 USD and slap a 16gb 9060xt (~25tflops @32bit compared to the PS5 gpus ~10tflops @32bit) in it for another ~350 USD it should theoretically do a better job - the GPU in the PS5 was on par with mid-high end PC GPUs on launch, but that was like 5 years ago.
- Quoted phrase in the opening sentence
- Opening quote: “phrase hash”
"It is still conceivable that these curves were selected to have a specific property in principle by searching over timestamps to the random number generator."
Yes.
"However, you would only need to do that if you were going to disclose the timestamp as some evidence that the seed was innocuously generated."
Does not logically follow when "doing and saying nothing" will completely suffice.
If you enjoyed reading about how you can approximate complex statistical values using basic programming, I highly recommend Jake Vanderplas "Statistics for Hackers" [0][1].
It converts the sometimes hard to grok statistics terms and formulas into some very simple loops and basic operations.
Can't recommend it enough.
0 - https://speakerdeck.com/pycon2016/jake-vanderplas-statistics...
One bad transmitter in your neighborhood can cut you off from the hobby. Someone who's into linux and computers can follow an online tutorial to rig together an SDR. Look at a bunch of numbers or static, shrug and move on to a different project without disconnecting that anttena breaking an entire band for all of his neighbors. Then it's on those neighbors to go build a directional antenna and buy portable radio gear and then do their own "in the field investigation" to isolate which house is doing that, then go try to explain to the random person they are polluting a band and hope the person is friendly and cooperative.
In a perfect world the FCC would do that, but getting the FCC to act on small infractions like that is a slow and long process and often goes unanswered. Especially if you're not near a marina or a federal building or an airport. So one hackaday reader can take down ham operation in an entire area because they were just interested in a weekend DIY project.
People throw that "against the law" to limit the damage. But actually getting the FCC to prosecute these issues is a challenge in itself.
Take this Hackaday article from the other day:
https://hackaday.com/2026/10/15/so-you-want-to-make-your-sdr...
The subject is mostly academic, but in 18 comments no less than 3 different people are warning about doing something against the rules. It's weird. I don't know of any other subject that provokes responses about "the rules" like this. I could understand if the article title was "How to Aim Laser Pointers at Airplanes", but attach a long enough wire to an oscillator and people start to get uppity.
On the other side of things, I don't know if you've turned on a shortwave radio recently, but the only thing I hear on mine is switch-mode power supply noise.
There is obvious large overlap. Both are nerdy and techy and attract similar base groups. But the ham space is dominated by people who want to be "good citizens" through collective action, while computer hacker is more about cyberpunk "if I could do it, then it's allowed".
Perhaps it's the clearest example of "The medium is the message" concept.
I also live in Denver, and while the quote is technically true, it is misleading. A Waymo did drive into a farmers market...before there were barricades up, when other cars were still driving through as well for set-up time. Nobody would have been charged with anything.
> “It entered through our food truck area,” said Nicole Jarman, the production director of the farmers market... there was a “Road Closed” sign on one side of the street, but the other side of the street was still open to vendor traffic during set up and the Waymo followed a vendor vehicle past the sign and then turned onto the 1400 block of South Pearl Street.
> “It’s still load-in, our barricades are moved to the side, so there’s vehicular access, and it followed that traffic, went slow, and then made its way on out at Arkansas and kept going” she said.
> “It was not anything that made us nervous,” Jarman said.
[0]: https://kdvr.com/news/local/waymo-drives-through-denver-farm...
Honestly the last thing I thought would happen would be that it would get lost. I thought maybe it would get confused on the road, but it drove better than most human drivers.
- Opening sentence uses a quote to challenge or distance a term
- Opening sentence challenges the term “bad”
As in "main ratings agencies" see no reason to bother
>However, Anthropic’s disclosure to KBRA excluded several important financial metrics, including free cash flow and earnings before interest, taxes, depreciation and amortisation. Credit rating analysts had to come up with their own estimates based on assumptions, the documents show.
>Given the limited information, KBRA produced a “risk score” to assess Anthropic’s creditworthiness rather than the traditional full financial models.
This and the fact they went to a small ratings agency does not inspire confidence in the rating.
I also remember the big claims about WebP's superiority, which ended-up in practice, at equivalent file-size, mostly trading away some of the JPEG artifacts that we all know for some unusual new ones.
I don't think there's a large-enough margin in this space that a new format could chew into to make the general public care: image compression seems like a solved problem overall. The second best thing is to come-up with a format that can make the content producers not have to care either, i.e. be the best in class or close-enough in a variety of use-cases (lossy, lossless, fast, animated, hdr, progressive, …). My naive impression was that jxl is mostly interesting in that regard.
It's the reason AI company leaders are so adamant about framing it as a post-singularity utopia with UBI. They have no answer other than "you're all going to be jobless, better figure that out." Worse yet, they see it merely as a policy problem, rather than a problem they're responsible for. It _is_ a policy problem, no doubt, but the angry mobs won't be coming for policy.
The only “disinflation” consumers have experienced is devaluation of products through obsolescence as newer, often more expensive alternatives enter the market.
More and more it’s seeming that the “societal benefits” from the last fifty years of progress have largely been concentrated at the top.
- Ordinary term used more than once in the same comment
- “covered well” appears 2 times
For example, the “social workers & care -> counsel parents with child rearing problems” category is considered “well covered” because there are 13 prompts.
It’s self-evident that the spectrum of work carried out by a social worker undertaking that work cannot be well represented by thirteen prompts.
I suspect if you really strictly defined “covered well”, to the extent that any normal person would use that phrase, the headline would be above 99.9%.
- Quoted phrase in the opening sentence
- Opening quote: “what are all those construction workers even doing?!”
Like any sizable population, there are absolutely some do-nothing inefficiencies to be culled. But most of the time it's just onlookers seeing a tiny slice of the day to be interpreted from their very limited vantage.
> ...
> But even after controlling for AI-exposure, O*NET tasks from occupations with larger wage bills are significantly overrepresented among benchmark prompts:
So, in short: they want AI to take the good jobs. AI is not coming for jobs with tasks like "gather and empty trash" or "service, clean, or supply restrooms," or "mix water and detergents or acids in containers to prepare cleaning solutions according to specifications." We software engineers and lawyers would be smart to start retraining (and adjusting our living standards) now!
And this points to one of the ways AI will help improve society and living standards for all: we'll have cleaner, better-maintained bathrooms, as labor one wasted on engaging intellectual tasks is redirected towards menial, low-pay tasks that still need a human touch.
Reason to be optimistic, amirite?
- Quoted phrase in the opening sentence
- Opening quote: “stakeholders”
If anything, AI has made it more fun simply because it means not having to think as much about the drudgery of details: syntax, API schemas, boilerplate code... And by "more fun" I mean "somewhat less unfun". Cause it's still not fun. But AI makes my burnout somewhat more tolerable.
I was someone who got into tech because they enjoyed coding as a mental exercise, not because they wanted to get rich. I feel like the fun part of the job has been now been automated away. All that remains is the boring part that I previously tolerated as a necessary evil: aligning stakeholders, kissing management ass, obsessing over the performance review cycle.
I know that some people have the inverse relationship: they find the "boring" part to be energizing and the "fun" part to be tedious. And I've seen a couple coworkers experiment with using Claude to agentically caress their boss's derriere. That's good for them and I wish them all the best. As for me, hopefully I can last long enough to engineer an early retirement. Then maybe I'll contribute to open source, assuming that there still is a FOSS community left by then.
It's always some variation of: the LLMs are incredible and gaining logical and reasoning ability at an astounding rate, but fortunately they will always need me to sit here to hit the "Enter" key.
I envy their optimism.
I feel like there's been a growing, dark underbelly of gambling addiction since it was legalized online. And it's going to be difficult to walk it back.
(the ones that truly blow my mind are the promo ads for people gambling on buying mystery collectible cards then virtually "ripping" the pack open. Grown adult men paying hundreds of dollars! Absolutely wild.)
- Ordinary term used more than once in the same comment
- “WAS” appears 2 times
This is my biggest problem with opus 5.5.
It's absolutely incredible in many ways, but it has no sense of "good enough".
It will go around in circles polishing something so to resolve edge-cases on edge-cases.
"Someone could create a class that calls your public method in such a way, that under 3 other race-condition properties, there might be stale data read".
It has no sense that it isn't appropriate to add a half dozen classes just to close that kind of hole, especially when there isn't even such a caller or possible conditions at present.
- Quoted phrase in the opening sentence
- Opening quote: “I Love GL_NEAREST_MIPMAP_NEAREST”
> "3D cards smoothed textures with bilinear filtering"
You can just not do that.
I'm guessing this is a joke and I'm just taking it too seriously, but the only thing that "proves" is that Windows lies to you lol. It's obviously information that's stored _somewhere_.
‘Larry, throw the cat at the keyboard!’
MEEEOW!
‘What did it type?’
‘&}Ak'>v=4%Qe}9;’
‘OK, so let’s make this a syntax for hello world.’
> We stopped when we got a clean compile on the following syntax:
for(;P("\n"),R--;P("|"))for(e=C;e--;P("_"+(*u++/8)%2))P("| "+(*u/4)%2);They recognise the faces, voices, unique gait, posture and body language of people from their local area, and can remember an average of around 30 people. Some have been known to recognise up to 100 different people. They remember people for around 5 years. They'll know who lives where, what your regular schedule is, and which path you regularly travel along. They even know all the local cats and dogs and even can identify specific bicycles, cars and bright clothing uniforms (such as postal uniform).
During breeding season, they are most famous for swooping anyone near their nest who they deem as a threat (about 10% of males swoop). They will avoid swooping people they recognise though, so friends get a free pass.
They also hold grudges, so if you do something to annoy or threaten them, they will remember you. Some people have been on a magpie's "enemies list", and returned to a location 10+ years later and immediately been targeted, as the bird recognised them. Some birds even teach their young which people are to be treated as enemies.
Not that unique. My grandfather was a forester, his favourite time passing was to whistle back bird's song, changing only a little bit at a time, turning it slowly into something totally different. The changed song was passed to other birds and offsprings no probs.
The result was quite hilarious, you went to the primeval woods and you could hear wild birds singing the Für Elise... I bet it freaked out ornithologists!
(My granddad died about 30 years ago, sadly the bird "dialect" has vanished since, but I still have some old tapes in the attic.)
- Quoted phrase in the opening sentence
- Opening quote: “small”
This is more about beehaiviour research and playing with your balls, but the basic idea seems to be that even "small" brains do the same stuff the big brains do - like forming relationships, playing for fun, using tools, etc.
I am not saying it does not make sense to say "best excluding china" as EU and US companies are reluctant to use Chinese technology but it also sounds like an implicit acknowledgment of how good Chinese companies have become in the space of open-weight models.
> If the Licensee or any of its affiliates provide a third party with access to artificial intelligence models as a product or service, via API or a similar interface, where the third party receives outputs of such models (“Model as a Service”), and the aggregate gross revenue of the Licensee and its affiliates exceeds twenty (20) million US dollars (or the equivalent in other currencies) during any consecutive twelve (12) month period, the Licensee must notify Meta and enter into a separate written agreement with Meta before offering the Software or any Derivatives as part of that Model as a Service business. This Section 3 does not apply to (a) internal use, research, or any commercial activity that is not Model as a Service, or (b) the Licensee's use of the Software as a component to power features of the Licensee's own product or service that is not a Model as a Service.
They're allowing the use of this in commercial products, just not selling the model directly.
And based on what I'm seeing, it's far from even being in that category as a service yet. Seems like the OP is conflating seeing a bunch of Muse ads on Facebook with its presence as a competing frontier model.
The nearest analogue would be "weights available", I guess.
- Quoted phrase in the opening sentence
- Opening quote: “very sticky”
"A minute has 60 seconds" is simple.
"A minute has 59, 60, or 61 seconds. Consult CGPM publications or your national timekeeping authority to find out." is not simple at all.
The page lists several good reasons for this, including:
- It's a common source of bugs.
- Solar time is not that important in a world of coarse timezones.
- IoT and anything that cannot receive time updates have trouble with leap seconds.
- Negative leap seconds have never occurred before, but might occur in the future, and almost no software has tested this case.
I'd also add:
- Unix timestamps are actually defined as "seconds since epoch minus however many leap seconds have happened". It's a hack to keep code that assumes "86400 seconds in a day" from breaking. It means that during a leap second the Unix timestamp goes back and repeats a previous value, making them ambiguous and non-monotonic.
- Leap second unpredictability can make it impossible to calculate how many seconds are between two UTC dates or Unix timestamps, for example the countdown to the year 2038 bug.
- The proper ISO8601 formatting of leap seconds ("23:59:60") is widely unsupported, breaking parsers, along with the assumption of how many seconds are in a minute, hour, or day.
---
EDIT: I've emailed hn@ycombinator.com to change the title. It's avoiding editorialization, but I think it's also preventing the post from getting the attention it deserves.
I'm booking a yearly one-second-long holiday on my calendar, to celebrate all the headaches this will save me. Happy end of leap seconds!
If you change your perspective from “I’m not gonna spend 500 hours on this” to “30 hours of great gameplay for 30 bucks” maybe it isn’t so bad.
I've warned users of many such instances of false claims, and safety issues with products. As a consequence, I've been banned from reviewing, and all my past reviews have been removed.
"Community standards", I guess.
Doesn't matter to me, neither their POS blood pressure cuff nor their scale will sync to HealthKit anyway (oh, how I've tried), I'm not losing anything. So stomping my feet with "I'm never buying another Withings product!" won't make any difference, I was not going to buy their POS products again anyway. :-)
- Quoted phrase in the opening sentence
- Opening quote: “do not test their builds with any of their allegedly 'optional' features disabled”
Much more in the LWN article, I'm not going to quote everything.
- Quoted phrase in the opening sentence
- Opening quote: “Will consent”
For AI, there were top down directives to find new ways to experiment with and incorporate into core products and processes, but as you went down the management chain, management became more conservative. First and second line managers were strongly incentivized to keep their products, roadmaps, and interfaces stable and predictable by similarly conservative stakeholders. And as such, they were generally unwilling to officially dedicate much effort to AI other than the minimum “look at how much we’re using Claude”.
In particular due to the lack of central tooling, any team trying to do much with AI essentially had to reinvent much of the stack themselves. So even small projects were largely untenable due to inevitable required scope creep.
But isn't the point that the neural nets learn some insanely abstract ideas anyways? So whether the complexity is "ahh we think this neuron does 100 seemingly unrelated things but we don't understand it" or "this actually is some kind of latent shared feature, but we don't know what the concept represents because the composition of the entire circuit is an arcane algorithm", we're kind of stuck, no?
This seems like a classic model-description issue more than anything.
The idea is: a model may only have 100M neurons or whatever. But there are infinite concepts in the world. Therefore, models have to "stuff" multiple concepts per neuron.
How do you do that? Theres two ways:
1. Each neuron handles similar concepts. One neuron for dogs, a separate neuron one for cars, etc.
2. Each neuron handles a bundle of unrelated concepts. One neuron handles dogs and cars. Another neuron handles cats and numbers. And so on.
Historically, people thought #2 was correct because many neurons were found that seemingly represented multiple concepts. #2 had a cool name (polysemanticity) and was promoted in OpenAI/Anthropic blogs so it had a lot of traction.
However, this blog says #2 is wrong. Actually, the dog/car neuron is just a "V shape neuron". So #2 was just methodological error, and #1 was right all along. This is all the blog is saying.
- Quoted phrase in the opening sentence
- Opening quote: “hostile to users”
They use this software to find and kill people. Which they then publicly brag about. How much more hostile do you want to be?
Like immagine if the rhetoric of Turkey was constantly talking about expanding to Greece and how the land of Greece is historically the Turks. Obviously that would make any Greek uneasy.
- Quoted phrase in the opening sentence
- Opening quote: “all you need is SQLite”
But I recently looked into self hosting and several things each wanted their own Postgres container. Immich wants a Postgres and a redis instance.
If you’re building apps targeted at self hosters, SQLite is really all you need.