It's hilarious that all of these people think they're writing like Shakespeare, or that they are Linus Torvalds in the IDE.
In reality, half of them are writing below average quality code and prose. Zero self awareness.
It's the same as the poster article on HN earlier. Most people couldn't do nearly as good, but criticize work that does the job sufficiently.
Me? I know I'm not the best coder or writer (or designer). AI helps me do things at a quality superior to what I could do without it. And I feel zero... ZERO shame about that.
Multiple jurisdictions have tried to play the stupid games of "we'll make a dumb fuck law that creates problems for you, the website operator, and if you don't conform, we'll sinkhole you at ISP level". Platforms like Xitter and Telegram being the common early targets.
The "age verification" push is a big move in that direction in Western countries. "Think of the children" is an evergreen excuse for authoritarian measures.
Ive been trialing different types of interviews for over 2 years now and here’s what I’ve found.
I tried 2 flavors of a technical interview and this round was 1 hour in total.
The first was a traditional interview problem where we give the candidate a codebase and a docker image to run against. The problem itself involves using the docker image to post and get responses from but the candidate is expected to write a new helper/service in any language they choose. We ask for no agentic workflows for this but any other resource is fair game. This has been the most successful for us to find candidates and generally provides the best experience.
Our full agentic interview involves an existing fullstack typescript codebase that we give to the candidate a week or two before. The candidate is free to use any agentic workflows they want during the interview and we give them a ticket that we went to implement during the interview. I find a ton of variability in this interview and a lot of people will just take the entire ticket and one shot into Claude without prepping. This is lead to the biggest disparity in results of candidates so we stopped doing it.
The story you are referring to involved the New York Post or its affiliates breaking in to Hunter Biden's laptop (a felony) and publishing his personal correspondence, complete with unredacted phone numbers and email addresses. This violated Twitter's policies against doxxing and dissemination of hacked materials. Characterizing this as censorship is misleading; the story contained no newsworthy information and its publication violated longstanding Twitter policy. It's hard to imagine Twitter functioning without such a policy. I don't see any reason the New York Post should get away with violating it just because it claims to be a newspaper.
The Wall Street Journal, a Murdoch rag like the Post and hardly known for its liberal bias, declined to publish this information, which is why it wound up in the disreputable Post instead. Respectable journalists have ethical standards. Act like thugs, get treated like thugs.
The part of this I always forget is how lazily the ranking is applied. HN does not rescore the list on every page load. A story is rescored when someone upvotes it, and separately one story out of the top 50 gets picked at random every thirty seconds and rescored. Pages cache for another ninety seconds on top of that.
So the front page is never a clean rendering of the formula. A story that stops getting votes can sit at the wrong position for minutes at a time.
Which is also why the penalty figures in the post are ranges rather than numbers. Shirriff is inferring them from stories ranked lower than their raw score implies, and a story that simply has not been rescored recently looks exactly the same from outside. He lists that as one of his sources of error, along with neighboring stories carrying their own penalties.
Instead of using nyaa[dot]si and dowloading the seasons that you want to watch, like a normal person would, you decide to pay for DRM content from time to time to get Crunchyroll's own little things and consider it a win for justice.
20 MHz is well within the range of any random scope, and you should have at least one "any random scope" on your desk if you're doing electronics seriously enough to reach the "diagnosing QA fail boards" stage. So probe the damn thing - and see if the crystal ever reaches a stable operating frequency once the board is powered. That would tell a lot.
And yes, the "holy book of SMD passives" is a must too. They cost you what, $40 each? And can save you days of waiting for the "right" passives to arrive when you need to test a circuit change. You don't need to have every size - footprints are negotiable when you're hand-soldering - but you should have at least one Book of Many Resistors and one Book of Many Capacitors.
systemd makes it so so easy, and integrates many many different mechanisms and systems for locking down what procesess can do and what they have access to.
"The brain does it" is not in itself a reason to do the same in AI.
Sometimes, what the brain does is genuinely a good solution to a given task - one that's good regardless of whether your neural network is of artificial or biological variety. But sometimes, what the brain does is an evolutionary kludge, or a hack that works around one of the "being made of flesh" issues - of which there are a great many.
We have known examples of both - and a few known features that might go either way.
The brain does have some known useful features that we are yet to plunder - usually because we know they're there somewhere but not how they work. We don't know how the brain stabilizes online learning, for example. Or what low k-complexity priors and data augmentation processes does it use to enable its sample efficiency.
But a two-system split? Useless by itself. Splitting a network in two is easy - but if we don't know what that split does, what it buys us, what useful bias does it impart? We're just adding complexity. See: the investigation into HRMs, and how the "hierarchical" part proved to be a lot less meaningful than anticipated.
I think basically AI posters are bad because when given freedom to the model (which is what people will do when just asking "make me a poster for my coffee shop"), the model will always tend to do the same design style. So yes, as this article says, with better prompting you CAN get better results. The other problem I think is tool calling, because when prompted for a poster, ChatGPT will definitely use image generation, when it should be generating some html and css + some image assets and then presenting the user an image or pdf.
I think the tendency to go for image generation always might be related to token costs, designing with Code & adding some generated images to the mix should be more expensive than just image generation.
This is US centric, because it's what I'm familiar with.
I'd think really hard about how the characters are coded. I haven't seen this particular show. Short skirt and ditzy is a cash grab to sell beer. Short skirts and extremely competent? that's a whole different thing. There are a million problems with the dukes of hazard. Daisy Duke is not one of them. Extremely competent character. Echoed by Marisa Tomei in my Cousin Vinny. the us has a whole other suite of issues around race, that I'm going to avoid. Wonder Woman falls into the same category I think.
The lesson, repeatedly, don't judge a book by its cover. at the time, it's funny, but it's gently poking fun at the audience for their sexism.
maybe it's some dumb campy show. maybe there's something more subtle that's being pushed, that's only really possible to push in science fiction. this isn't us, this is some far-away future.
Gloria Steinem continued to wear mini skirts. a playboy bunny is a ditzy girl with no value beyond delivering drinks, which is gross. Legally blonde fits into the category as well, you can't just look at the surface, you have to look at the whole person.
I can't speak to this particular show. but maybe try to put yourself into the mindset of that time, and see if they're showing you something about that worldview.
AlphaStar beat one retired professional by cheating.
AlphaStar won a showmatch against TLO. TLO was never one of the strongest players in the world. He had been retired for over three years by the time of the match. Google set the rule that their system would have human-like mechanics, but it played several times faster than any human, never issued a wasted action, had an inhumanly fast reaction time, issued commands with perfect accuracy using an API, and could see the entire map at once.
It was later released to the open ladder with more human-level mechanics. Even strong amateurs regularly beat it. I have beaten it myself. It was strong, but not even close to the level of the strongest human players. It had obvious and easily-exploitable deficiencies in strategy and building placement.
I think even the cheater version would have lost handily to Serral or any of the strongest players.
(It apparently beat MaNa as well as TLO, but those matches were never released to my knowledge. I see no reason to assume Google cheated less flagrantly in private than they did in public.)
> it’s reading out your desired specs to a team member who does the actual engineering work
Congratulations, you just discovered ”delegating”. I hope you don’t have anyone reporting to you at the moment because if this simple concept sounds demeaning to you… wow.
I don’t think there’s ever been a more accurate and appropriate moment to say: that sounds like a you problem.
If you think computer science is only about manually writing lines of code yourself then goodness me you have SO MUCH to learn.
> Was this, then, done to impress others rather than enrich the coders skills and experience?
So coding is only valid when it “enriches the coders skills and experience”? How do you define those? Better yet, WHO is defining those? Are you so narrow minded and decidedly contrarian that you simply refuse to think of any other reason the dev might’ve done this for? I don’t wanna tell you you need to touch grass but goodness me you really ought to consider opening your mind to the fact that you are not the centre of the universe.
> The futuristic equivalent of pressing the “demo” button on old electronic keyboard and pretending to play
These two things do not equate in the slightest and the fact that you think they do is utterly depressing.
A lot of things are like that though. There's a cartoon - I thought XKCD but I can't find it - about how a beginner might identify an oak leaf by a set of descriptions - but an experienced arboriculturist would just say "well, you know, it's an oaky shape".
Like, I can tell a crow from a blackbird. Could I explain to you how to tell a crow from a blackbird? Absolutely not.
This absolutely doesn't look like AI writing and this is a pretty weird signal to use.
I have a much better theory: as it happens, Substack automatically converts "--" to em-dash as you type. I'm guessing the author knew that and wanted to use this feature, but composed the article in another editor and then copied-and-pasted the whole thing. Because of how Substack implements the feature, the substitution doesn't actually kick in on Ctrl-V, so all these "--" remained in place.
So i wanted to know what an Android phone actually transmits when you aren't touching it most articles say "Google collects a lot of data" but nobody knows to what extent and publishes some proof lets say some packet captures
Stock Android 16 averaged 348.4 requests/hour to Alphabet ASN 15169 endpoints turning off location and usage diagnostics in settings still left 194.2 requests/hour active mostly checkin.gstatic.com Wi-Fi BSSID surveys. Running GrapheneOS with Sandboxed Play Services dropped traffic to 12.1 requests/hour, while pure GrapheneOS registered zero.
It’s possible that they’re hiding that the friendly fire incident in Kuwait in March 2026 (3 jets destroyed, $250mm+ loss) was a real incident caused by AI. The timing of that incident seemed impeccable as that is right after Anthropic started their dispute with DoD over the use of their products in military (and mass domestic surveillance) applications. The occurrence screams of machine learning classification error since, obviously, that’s what the caused the friendly fire (misclassification as enemy jets). Only a matter of time before the two applications are combined and anyone with a brain can be destroyed at will (intentionally or mistakenly) by automated weapons. This is what citizens of this world should really be fearing. (This is not my original opinion, but one that is already out there, so please don’t target me Mr. Ai).
It's hilarious that all of these people think they're writing like Shakespeare, or that they are Linus Torvalds in the IDE.
In reality, half of them are writing below average quality code and prose. Zero self awareness.
It's the same as the poster article on HN earlier. Most people couldn't do nearly as good, but criticize work that does the job sufficiently.
Me? I know I'm not the best coder or writer (or designer). AI helps me do things at a quality superior to what I could do without it. And I feel zero... ZERO shame about that.