HNHacker News
TopNewBestAskShowJobs

grues-dinner

1,956 karma · joined July 24, 2024

submissionscomments
grues-dinner··on Progress towards universal Copy/Paste shortcuts on Linux
Same in Firefox. Absolutely infuriating and trivial for them to fix/disable with an about:config even if you absolutely had to keep Ctrl-Shift-C for the masses. I suppose it only affects, oh, everyone in the world who uses both dev tools and a terminal. Which is almost a complete intersection with people who use dev tools! Or have they always been expecting AI to take over web design entirely and you no longer need terminals or something?

You can inject JS into every single page to fix it via an extension, but extensions don't have the ability to fix it otherwise. I suppose you could also intercept it in an IME?

grues-dinner··on Launch HN: Golpo (YC S25) – AI-generated explainer videos
It seems really strange that you wouldn't farm this kind of thing out to a non-AI function that animates the text properly into the space given using parameters that the AI generated. I mean it's impressive that it does work at all, let alone as well as it does, but are we also going to get an AI to do all the video encoding as well?
grues-dinner··on Show HN: Keeps – Mail a postcard that plays your voice
It's definitely an attack vector we'll be seeing more of. My bank uses voiceprints as a password, for example.

My bet is some of the big "we record this conversation for quality and training purposes" databases get stolen and deanonymised for a mass leak of millions of voiceprints.

I currently have a rule with my parents that only I'm allowed remote access to their computer of any kind. And if anyone else wants access they should ask me first. Maybe one day I'll have to add some "nuclear biscuit" pre-shared code cards to that.

grues-dinner··on Show HN: Keeps – Mail a postcard that plays your voice
I thought (from the title) it was going to be a service where the audio was downloaded into a physical sound chip like a musical birthday card.
grues-dinner··on How to teach your kids to play poker: Start with one card
I've only played at home and everyone announces the tiles (so an interruption every 5 seconds) so maybe that's the problem. Maybe if you're good enough at it you can do the game on autopilot while chatting? I'm always slowest even to stack the tiles at the start and that's without distractions!
grues-dinner··on How to teach your kids to play poker: Start with one card
It's a different kind of skill that is more about predicting what your opponent will do based on the same information at which you are both looking.

Chess and go even more so are perfect information games, but there is substantial risk in strategies than can be derailed by the opponent noticing them too early or even by not noticing and ignoring bait.

grues-dinner··on How to teach your kids to play poker: Start with one card
Careful what you wish for! Mahjong is the opposite: you're always playing or setting up the next round, there's no down time, you can't stop paying attention even for a moment or you might miss an important tile, and you can't even skip a round for a comfort break.

Also it seems to be complex enough there is no mental space or time left to talk about anything other than the tiles. Exhausting!

grues-dinner··on Pricing Pages – A Curated Gallery of Pricing Page Designs
I would be except that the feeling seems to be that you get screwed either way:

* Tiers (aka new car model): something is always strategically left out of the otherwise "ideal" tier to force you up a level, even though you won't use most of the other options. Sometimes the "nearly there" tier is artificially expensive to drive you to the higher tier - the same trick as a medium coffee being only fractionally cheaper than the large. Sometimes there's a ratchet where you can upgrade but a downgrade is a huge hassle and/or penalised.

* A la carte (aka the car/dishwasher spares model): every option feels expensive and you feel like you're being nickel-and-dimed and you know the marginal cost of providing that option was small

* Top-up (aka the phone minutes model): top ups are obscenely expensive and are either a punishment for being "cheap" (i.e. prudent) or act as a threat to push you up a tier in the first place

Add a few special offers, points, cost sinks and lock-ins (especially where hardware is involved), rewards and all that crap here and there to muddy it up to prevent a clear comparison being made. I basically assume all subscriptions are doing some kind of mind-games or scam with every little aspect of the pricing.

Not that a fair price can't be any of the above options. The vendor has to cover the overheads somewhere!

grues-dinner··on The surprise deprecation of GPT-4o for ChatGPT consumers
> Even ChatGPT 5 confirmed this,

>> why does the gpt-4o api not do this?

> ChatGPT said:

>> Because the GPT-4o API is tuned and delivered in a neutral, low-intrusion style by default.

But how sure are you that GPT-5 even had this data, and if it has it, it's accurate? This isn't information OpenAI has publicly divulged and it's ingested from scraped data, so either OpenAI told it what to say in this case, or it's making it up.

grues-dinner··on AI Ethics is being narrowed on purpose, like privacy was
Yes, all this highly public hand-wringing about "alignment" framed in terms of "but if our AI becomes God, will it be nice to us" is annoying. It feels like it's mostly a combination of things. Firstly, by play-acting that your model could become God, you install FOMO in investors who see themselves not being on the hyper-lucrative "we literally own God as ascend to become its archangels" boat. You look like you're taking ethics seriously and that deflects regulatory and media interest. And, it's a bit of fun sci-fi self-pleasure for the true believers.

What the deflection is away from is that the actual business plan here is the same one tech has been doing for a decade: welding every flow and store of data in the world to their pipelines, mining every scrap of information that passes through and giving themselves the ability to shape the global information landscape, and then sell that ability to the highest bidders.

The difference with "AI" is that they finally have a way to convince people to hand over all the data.

grues-dinner··on VIN: The 17-character code that runs the automotive world
I think that is extremely common in adverts (most famously here, the M&S "It's not just bread. It's our stone-baked, hand-shaped artisanally-molested bread"), and narrative media like that distinctive kind of journalism written like a storybook, and which (I think) then bled into popular media like true-crime podcasts: "It wasn't just Tuesday. It was the last Tuesday he would ever see. <intro music>".

Also this kind of short sentence construction is used in the incredibly annoying and pervasive style of headlines for opinion pieces: X is Y. And it's Z. (where Z is often "not OK" or "OK").

I assume all this overuse is where LLMs picked it up and weighted it highly.

grues-dinner··on Project Hyperion: Interstellar ship design competition
I Blame Your Mother
grues-dinner··on Teacher AI use is already out of control and it's not ok
Even then, Slaughterbots did it first: https://youtu.be/9fa9lVwHHqg?t=129

And you probably don't need fiber optic, because you're operating in "owned space" - the drones sit on charging platforms until needed. You can have, say, additional access points embedded in the walls and ceilings (for a price, but it's children's safety, so who are we to say base station rental is worth more that little Timmy not getting a 5.56 in the back in a signal blackspot!?)

grues-dinner··on Teacher AI use is already out of control and it's not ok
I foresee lots of up-sells:

* Class monitoring when no teacher present - optional collusion analysis

* Conduct enforcement in corridors - optional RFID speed ticketing to prevent running in the hallways!

* Playground overwatch - optional score keeping for licensed games such as Hopscotch [TM].

* Perimeter monitoring for truants, contraband trading and drug dealing

* Toilet break escorting (optionally at a discreet distance)

* Per-student tracking and ensemble fraternisation analysis, optional social media and online profile correlation, and real-time alerting of parental accounts on contact with other students in parent- or community-provided watchlists or handy pre-set demographic groups.

* Student mood, wellness and attitude monitoring based on body language and speech patterns. Referral to preferred behavioral therapeutic partner providers at a discount!

With facial recognition you can even send warnings and punishments directly to the student and parental phones via the CGA App and apply demerits to their account automatically. Link a lunch payment account for automatic profanity penalties!

grues-dinner··on Is the interstellar object 3I/ATLAS alien technology? [pdf]
I'm here for all of that that.

Using the Wikipedia example of a 1km, 5600g superconducting rail gun that launches at about 10km/s, we just need about a 10km gun to achieve 30km/s (length goes with launch speed squared).

Put it on the lunar surface for a roughly 2.5km/s penalty (I think, plus you obviously need to shoot it when the moon faces the right way).

No humans though, far too squishy. But you could launch a whole swarm of microprobes which could be a very effective distributed observation platform with a gigantic baseline.

If you haven't read it, the short story Maelstrom II by Arthur C. Clarke has a lunar rail gun in it. And the rest of the The Wind from the Sun collection is very good. The namesake solar sailing regatta story is great too.

grues-dinner··on Teacher AI use is already out of control and it's not ok
I thought it was the other way around, the guard is there to shoot the dog when it's attacking the children (with apologies to the very old joke about catching bears in trees).
grues-dinner··on Teacher AI use is already out of control and it's not ok
> The drones can also fly through windows, using a front lance to break through.

https://www.campusguardianangel.com/faq

I would say you couldn't make it up, but you could. You'd just be called a bad writer with unsubtle and derivative ideas.

grues-dinner··on Teacher AI use is already out of control and it's not ok
Drone-based school security with flash bangs, pepper spray and physical dive-bombing of the attacker is already the plan: https://www.campusguardianangel.com/

Just add a student compliance add-on subscription.

grues-dinner··on Scientists shine a laser through a human head
It's interesting to me that hair or skin colour make such a difference when it's only a mm or two out of a far thicker solid lump of brain, bone and so on.
grues-dinner··on Is the interstellar object 3I/ATLAS alien technology? [pdf]
ATLAS was expanded in 2022, and this object was discovered by the new telescope in Chile - the more we look the more we find. With more survey telescopes coming online (Vera Rubin just recently, Nancy Grace Roman and Xuntian coming soon to name a few), I suspect we'll start seeing quite a few more of these.

Hopefully we humans get a mission ready to go that will allow to go and have a look when a suitable one turns up with enough notice. Presumably one that isn't nailing though quite so fast as 3I/ATLAS (ʻOumuamua and Borisov were about half the speed each - about 30km/s). Annoyingly the speeds mean that really all you can do is a very fast flyby, unless you are incredibly lucky with trajectories, the object moves very slowly or we can ship a truly massive amount of "rapid" (e.g. not ion engines if you want to catch it this side of the heliopause) delta-v to orbit.

The rocket equation is really not on our side here if we wanted use chemical means. If you have a specific impulse of 300 seconds, you basically cannot get a 100kg probe to 30km/s delta-v without a slingshot. 100 tonnes of fuel gets you to about 20km/s, 2000 tonnes gets you to 30km/s. And a craft that holds 2000 tonnes of fuel probably masses more than 100kg.

Maybe the better bet is a really good sunshield and then everyone works on their cardiac health so they can see the intercept in 30 year's time, and even then it's a blink-and-you-miss-it flyby at over 20km/s: http://orbitsimulator.com/BA/lyra.gif

grues-dinner··on The US military’s on-base slot machines
Gambling is an evil business and everyone willingly involved in peddling it should be exiled from society one way or the other. FOBTs are especially streamlined. Everyone in the business knows exactly what they are doing.

That said, the same applies to lots of things: drugs almost literally, but also porn and porn-adjacent services are obvious brain-hacks, but also food, drink and even shopping are frequently structured in such a way to capture business through non-consensual manipulation of the human mind.

None of it is new: people have been losing their shirts, homes, families and lives to most of the above since before money existed, but technology-wielding capital isn't exactly freeing us from it, but instead has optimised it. Sure, it's freedom, but it's freedom of both predator and prey.

grues-dinner··on ThinkBook Plus Gen 6 – Lenovo’s rollable laptop
Hey! All these devices will still work just fine if I can only figure out how to put them back together again. Except all the ones with the broken plastic clips.
grues-dinner··on So you want to parse a PDF?
The docx standard, which was rather tendentiously named Office Open XML back when OpenOffice was still called that, is 5000 page long and that's only Part 1 of ECMA-376, with another 1500 pages of "Transitional OOXML" in Part 4 which is basically Word-specific quirks.
grues-dinner··on Job-seekers are dodging AI interviewers
"Societies" (YC25) seems like a good first step! https://news.ycombinator.com/item?id=44755654
grues-dinner··on OpenAI's "Study Mode" and the risks of flattery
Embedding "your" AI at every level of everyone else's education systems seems like the setup for a flawless cultural victory in a particularly ham-fisted sci-fi allegory.

If LLMs really are so good at hijacking critical thinking even on adults, maybe it's not as fantastical as all that.

grues-dinner··on OpenAI's "Study Mode" and the risks of flattery
> Akin's Law #19: The odds are greatly against you being immensely smarter than everyone else in the field. If your analysis says your terminal velocity is twice the speed of light, you may have invented warp drive, but the chances are a lot better that you've screwed up.
grues-dinner··on Show HN: Print the daily weather forecast on a thermal receipt printer
The main theoretical risk is to workers who handle receipts often at work. A checkout operator could have hundreds of times normal exposure and checkout workers were reported to have high (3x baseline) BPA urine levels in some studies (but not others).

It's the same for lead free solder. Occasional amateur exposure is probably not a major concern, and especially not for adults who won't get pregnant.

Then again, if you're making a thermal printer device for fun that you plan to touch daily, especially if you want to have children using it, who are also likely to touch it with more of their hands and/or get it wet and/or put it or their hands in their mouths, loading the device with phenol-free paper seems sensible, practical and not especially onerous.

grues-dinner··on OpenAI raises $8.3B at $300B valuation
I assume the plan is getting AI into everything and making millions of data centipedes that eventually lead back to their APIs.

Then embed ads.

grues-dinner··on “No tax on tips” is an industry plant
> In the UK at least this is still very much the case.

Except for the very pervasive 12.5% "discretionary" service charge the you would have to request to be removed. Which is a genius piece of social engineering.

grues-dinner··on “Good first issues” are usually not good first issues
Seriously, every project needs to have the "inner dev loop" in the README as bullet points. Nothing sucks more than taking two days to figure out how to even build, and once built, how to actually install/run/debug it. Especially for projects that use a fleet of repos that all interact or involve multiple libraries that need to be carefully orchestrated to actually run.
← PreviousPage 8 of 22Next →