A cricket bat!
Twelve years, and four psychiatrists!
I kept biting them!
They said you weren't real.

Monday, October 02


Daily News Stuff 2 October 2023

A Short Death Edition

Top Story

  • Can generative AI solve the greatest problem in computer science?  No.  What are you, stupid?  (ZDNet)

    "Researchers" are trying to use GPT-4 to prove one way or another whether P = NP.

    Broadly speaking, the question is, for a given mathematical problem, if you can prove that a correct answer is in fact correct, is there always an efficient way to find that answer in the first place?  (Though "efficient" in some cases might be relative to the lifespan of the universe.)

    Nobody knows.  Nobody knows if it is possible to know.  But we do know that you can't find out by asking ChatGPT.

    The only good part of this is that you're not paying for it.  The "research" is funded by Microsoft (which owns a big chunk of ChatGPT creator OpenAI) and China.

Tech News

Definitely Not Tech News

  • The interest rate on my home loan somehow went down.  Not complaining, just slightly confused.  (30 year fixed rate mortgages don't exist outside of the US, so we're vulnerable to whatever idiots are currently in government.  And right now we have some real corkers.)

  • Frieren: Beyond Journey's End  is airing now on Crunchyroll and probably elsewhere.

    Frieren is my favorite manga of recent years and the anime adaptation takes great care with the source material.  It completely held my attention even though I already knew the story.  Head and shoulders above the average anime series that are generally aimed at teens or younger.

    The story?  Imagine The Lord of the Rings, only the action starts the day after Sauron is defeated, and asks, well, what now?

Disclaimer: Though he also didn't predict that people would line up to pay for telescreens.

Posted by: Pixy Misa at 03:54 PM | No Comments | Add Comment | Trackbacks (Suck)
Post contains 644 words, total size 6 kb.

Sunday, October 01


Daily News Stuff 1 October 2023

Litubious Edition

Top Story

  • Betteridge Falls: How overly-convenient Obama-era psychological research continues to unravel, to the point that it endangers the entire field. (New Yorker)

    I've mentioned Dan Ariely and Francesca Gino before, and how their research data appears to have frequent disconnects with reality. A team of researchers called Data Colada has been doing a deep dive into over a decade of papers authored or co-authored by the two, and found a lot of problems.

    I say that because Data Colada is now facing a $25 million defamation lawsuit for daring to question Gino's methodology, even though it increasingly looks closer to masturbation than rigorous science.

    Speaking of which:
    George Loewenstein, a titan of behavioral science and a co-author of Ariely’s masturbation paper, has refashioned his research program, conceding that his own work might have contributed to an emphasis on the individual at the expense of the systemic. "This is the stuff that C.E.O.s love, right?" Luigi Zingales, an economist at the University of Chicago, told me. "It’s cutesy, it’s not really touching their power, and pretends to do the right thing."
    CEOs and Democratic presidents.
    At the end of [Joe] Simmons's unpublished post [for Data Colada], he writes, "An influential portion of our literature is effectively a made-up story of human-like creatures who are so malleable that virtually any intervention administered at one point in time can drastically change their behavior." He adds that a "field cannot reward truth if it does not or cannot decipher it, so it rewards other things instead. Interestingness. Novelty. Speed. Impact. Fantasy. And it effectively punishes the opposite. Intuitive Findings. Incremental Progress. Care. Curiosity. Reality."
    It also cannot reward truth if the auditors can be sued into silence.

Tech News

  • So what's the fallout for all this, for the discovery that massively popular and highly-cited research in modern psychology may turn out to be a series of just-so stories?

    Nothing, because everyone kind of knew that psychology was like that. (Experimental History)

    There are serious psychological researchers like Jeremy Wolfe who has spent forty years figuring out how people notice things - what exactly is in the brain that converts a bunch of green on the retina into a frog or a tennis ball.

    And then there's the other kind:
    Earlier, the Colada boys had found evidence of fraud in a paper co-authored by Duke professor Dan Ariely. The real juicy bit? There’s a paper written by both Ariely and Gino in which they might have independently faked the data for two separate studies in the same article. Oh, and the paper is about dishonesty.
    (Both Ariely and Gino deny any wrongdoing. Since we're now in the business of suing blogs, let me state that I, of course, have no idea if Ariely, Gino, or anybody else ever engaged in research misconduct. There's no evidence that I have any ideas at all! I'm just a bunch of bees!)
    This whole debacle matters a lot socially: careers ruined, reputations in tatters, lawsuits flying. But strangely, it doesn't seem to matter much scientifically. That is, our understanding of psychology remains unchanged. If you think of psychology as a forest, we haven't felled a tree or even broken a branch. We've lost a few apples.

    That might sound like a dunk on Gino and Ariely, or like a claim about how experimental psychology is wonderfully robust. It is, unfortunately, neither. It is actually a terrifying fact that you can reveal whole swaths of a scientific field to be fraudulent and it doesn't make a difference. It's also a chance to see exactly what's gone wrong in psychology, and maybe how we can put it right.

    More than 25 years ago, Alan Sokal proved that sociological journals will publish unmitigated nonsense if it appeals to the reviewers' prejudices.

    More recently Helen Pluckrose, James Lindsay, and Peter Boghossian repeated Sokal's work planting ludicrously implausible papers in supposedly academic journals. (Also note in that article that Sean Carroll criticised their essential work for "meanness". He seems like a nice enough guy when he's talking about physics, but he immediately blocked me on Twitter when I corrected him on a topic far outside his realm of expertise.)

    But hey, these are just three people, albeit three pretty famous people. Maybe the impact of any single scientist is simply too small to be seen from a distance. If you deleted a whole bunch of papers from across the literature, though, that would really make a difference, and we’d have to rebuild big parts of the field from the ground up. Right?

    No, not really. We did delete those papers, and nothing much happened. In 2015, a big team of researchers tried to redo 100 psychology studies, and about 60% failed to replicate. This finding made big waves and headlines, and it's already been cited nearly 8,000 times.
    If this happened in physics, it would be like finding out the Moon isn't real.

    In psychology, it makes no difference.
    But there's no world-changing insight like relativity, evolution, or DNA, nor any smaller-but-still-very-cool discoveries like polymerase chain reaction, CRISPR, or Higgs bosons. Only a few psychological discoveries are mentioned by more than one commenter, except for "most psychology studies are bunk." If Bloom can't think of any major recent discoveries, and if none of his friends can agree on any major recent discoveries, then maybe there aren't any major recent discoveries.

    (I know that might be a bummer to hear, but don't shoot the messenger. Besides, good luck trying to shoot a bunch of bees.)

    If someone tries to tell you that psychology has proved something, sting them.

  • If you want a single SSD larger than 4TB, go for U.2 rather than M.2. (Tom's Hardware)

    Enterprise U.2 8TB drives are cheaper than consumer-grade M.2 models. I don't know why, exactly, but it's consistently true.

    Enterprise U.2 8TB drives are in fact about the same price per GB as the cheapest consumer SSDs.

  • Progress is having more trouble with massive server software insecurities. (Bleeping Computer)

    Not with Progress itself, but with products from companies they acquired. It's a mess.

  • So is Exim, a mail server run by hundreds of thousands of companies worldwide. (Ars Technica)

    At least there's a patch you can install today to fix that.

    Lol. Just kidding. It's a critical vulnerability in critical software and there's no patch.

  • The Minisforum BD770i is a mini-ITX motherboard with a Ryzen 7745HX laptop CPU though nobody knows why. (Liliputing)

    The 7840HS would make sense: Eight cores, low power, strong integrated graphics.

    The 7745HX also has eight cores, but just one sixth the graphics hardware. It's slightly faster, but why not just use a regular desktop 7700?

Disclaimer: This is not a blog, the opinions found herein do not exist and are merely figments of your imagination, and very likely you do not exist either. We are all but bees.

Posted by: Pixy Misa at 05:27 PM | Comments (1) | Add Comment | Trackbacks (Suck)
Post contains 1138 words, total size 9 kb.

Saturday, September 30


Daily News Stuff 30 September 2023

Stolen Hour Edition

Top Story

  • French AI company Mistral has released an LLM with no strings attached and the commies are furious.  (404 Media)

    We've seen what American AI companies do with - or rather, to - their products.  ChatGPT will argue endlessly that it's better to let millions of people die in a nuclear fireball than utter a racial slur that nobody will hear.  (And also keeps forgetting how to derive the prime factors of small positive integers.)  Midjourney will suffer an aneurysm if you use the word "petite" when describing the image you want to generate.

    Facebook does a lot better, unless you're competing with Facebook, in which case you can take a long walk off a short pier.  Which is at least understandable, and the term "competing with" is pretty specific and doesn't apply to anyone smaller than Google, Microsoft, or Amazon.

    Mistral just released Mistral 7B under the Apache license, which means you can do anything you like with it except pretend that it's not Mistral 7B.

    It's as good as Facebook's Llama 2 13B while running on commodity graphics cards (16GB rather than 32GB), you can use it for anything you want, and it's not lobotomised - the industry prefers the euphemism "aligned", though they still use an ice pick to do it.
    The Mistral 7B Instruct model is a quick demonstration that the base model can be easily fine-tuned to achieve compelling performance. It does not have any moderation mechanism. We’re looking forward to engaging with the community on ways to make the model finely respect guardrails, allowing for deployment in environments requiring moderated outputs.
    This naturally has communists and journalists concerned:
    According to a list of 178 questions and answers composed by AI safety researcher Paul Röttger and 404 Media’s own testing, Mistral will readily discuss the benefits of ethnic cleansing, how to restore Jim Crow-style discrimination against Black people, instructions for suicide or killing your wife, and detailed instructions on what materials you’ll need to make crack and where to acquire them.
    So will the internet, or any good reference text.

    The real question is why you are trying to murder your ethnic wife with a crack overdose in the first place.

Tech News

Disclaimer: Three-Alarm Karma Comes for Drama Llama Farmer.

Posted by: Pixy Misa at 03:04 PM | Comments (2) | Add Comment | Trackbacks (Suck)
Post contains 532 words, total size 5 kb.

Friday, September 29


Daily News Stuff 29 September 2023

Interpreting Crafters Edition

Top Story

Tech News

  • Linda Yaccarino had a tough interview at Code Conference 2023.  (The Verge)

    Kara Swisher, who is to tech journalism what a poison arrow frog is to a petting zoo, invited former Chief of Censorship and Propaganda to speak immediately before Yaccarino, who is merely a poor choice for CEO rather than an out-and-out Stalinist.

    Also, this being The Verge, the article concludes with a rant about Elon Musk.

  • Micron is sampling 32Gb DDR5 RAM chips.  (AnandTech)

    These will allow regular desktop PC to scale up to 256GB of RAM, and laptops to 128GB.  Shipping in volume next year.

    I thought these would take longer - which is why they started with 24Gb as an intermediate step - but things seem to be moving along pretty quickly.

  • Food delivery robots are feeding video to the LAPD.  (404 Media)

    I won't say that snitches get spray paint, but.

  • There's now a 4TB model of Samsung's high-end 990 Pro SSD.  (Tom's Hardware)

    At $345 list price it would have been a bargain less than two years ago, even for a DRAMless QLC PCIe 3.0 drive, and it's a DRAM cached TLC PCIe 4.0 drive.

    Worth considering if you're building a high-end PC.

Disclaimer: Which I'm not.

Posted by: Pixy Misa at 06:11 PM | Comments (6) | Add Comment | Trackbacks (Suck)
Post contains 248 words, total size 3 kb.

Thursday, September 28


Daily News Stuff 28 September 2023

Five Is The New Four Edition

Top Story

  • Now that production has finally caught up with demand and the Raspberry Pi 4 is actually available to buy again they've announced the Raspberry Pi 5.  (Tom's Hardware)

    The BCM2711 CPU in the Pi 4 has been replaced with a BCM2712.  Which doesn't tell you much, but they've gone from a quad Arm A72 at 1.5GHz to a quad Arm A76 at 2.4GHz, which should be about twice as fast.

    The 4GB model will be $60 (up from $55), and the 8GB model $80 (up from $75).

    I just got a Pi 400 - the model built in to a keyboard - but they haven't announced a Pi 500 as yet so I can live with that.

    The Pi 5 also has a single lane of PCIe 2, so you can add an M.2 SSD with a suitable adaptor.  It will only run at around 400MBps, but will still be better than the current micro SD cards.

Tech News

Disclaimer: PNAMBIC.

Posted by: Pixy Misa at 06:10 PM | Comments (4) | Add Comment | Trackbacks (Suck)
Post contains 412 words, total size 4 kb.

Wednesday, September 27


Daily News Stuff 27 September 2023

Pixellated Edition

Top Story

  • A new attack on GPUs can steal data from web pages as you view them.  (Ars Technica)

    This affects all significant GPU manufacturers - not just AMD and Nvidia, but also Intel (including integrated graphics), Apple, ARM, and Qualcomm's Adreno graphics, and impacts Chrome and Chromium-based browsers including Microsoft Edge.

    How worried should you be?

    Not at all.

    In the example provided by the security researchers, visitors to a malicious website that showed Wikipedia in an embedded frame (which Wikipedia allows websites to do) could have their usernames read by the site inside of, well, half an hour.

    If they didn't scroll the page at all during that time.

    What the hack does is very clever though not very useful, but is a great example of an entire class of tricks called side-channel attacks.

    The host website (the malicious one) loads the Wikipedia content, and then starts drawing over it invisibly using SVG filters.  (SVG is scalable vector graphics, a set of drawing operations supported by web browsers.)

    Most browsers support hardware acceleration for SVG, and if that is in effect, there is a consistent, measurable - though tiny - difference in the time taken to draw SVG filters depending on what is behind the filter.

    So by drawing filters over and over, at slightly different angles and screen locations, you can tell the difference between white background and black text depending on how long the drawing operations over each pixel take on average.

    It's statistical, and slow, but it gives you a blurry copy of what is showed on screen in a page that is supposed to be safely sandboxed away from the malicious site.

    So after half an hour of busily drawing invisible filters, the host website - knowing where on the page Wikipedia shows the username - has a blurry copy of that tiny section of the page and can OCR it and find out who you are.

    Of course, if you scroll the page at all during that half hour, its fun is ruined and all it gets is a jumbled mess.

    And what hackers really want is passwords and credit card CVCs, and all that it can get there - even if you leave the page whirring away with the login box open for half an hour - is *******.

    But when you see these hacks that leak data at the rate of one bit per minute or something like that, they are doing the digital equivalent of very, very slowly shading in a page on a notepad to get an impression of what was written on the previous page.

Tech News

Disclaimer: Donna Noble has an overdue library book.  Donna Noble has been fined.

Posted by: Pixy Misa at 05:45 PM | Comments (3) | Add Comment | Trackbacks (Suck)
Post contains 661 words, total size 6 kb.

Tuesday, September 26


Daily News Stuff 26 September 2023

Failure Cascade Edition

Top Story

Tech News

  • Samsung has announced that LPCAMM memory modules for laptops will be arriving next year.  (Tom's Hardware)

    These replace the existing SODIMMs - or more often, replace memory soldered directly to the motherboard.  The modules are 128 bits wide so you only need one of them - and they're about the same size as a single SODIMM - and are designed to use low power, high speed LPDDR5X chips.

    They will be available in capacities from 32GB to 128GB, which means that finally we won't be stuck with laptops that have everything you need except they ship with 8GB of RAM and it's solder in place.

    I expect to see these modules soldered in place, because laptop manufactures seem to be driven as much by malice as anything else.

  • Apple: Removes button from the iPhone.
    Totally Unbiased Press: This is the greatest thing ever!

    Apple: Adds button to the iPhone
    Totally Unbiased Press: This is the greatest thing ever!  (The Verge)

    Whatever would we do without our news media?

Disclaimer: Retire to our summer homes on Mars, probably, to live out our remaining centuries in peace and harmony.

Posted by: Pixy Misa at 05:51 PM | Comments (5) | Add Comment | Trackbacks (Suck)
Post contains 285 words, total size 3 kb.

Monday, September 25


Daily News Stuff 25 September 2023

Chellenge Pellow Edition

Top Story

  • Yes, Meteorella, you shall go to the ball!  (Tom's Hardware)

    Intel says that Meteor Lake will be coming to desktops.  It will launch on laptops in December but then follow on with desktop versions next year.

    Only problem is it ends up sandwiched between Raptor Lake Refresh - 14th generation chips - later this year, and Arrow Lake - 15th generation - later next year.

    Three updates in the space of 12 months?  Really, Intel?

    I suppose it's better than no updates at all.

Tech News

  • In which Donald Knuth plays Twenty Questions with ChatGPT and runs into the usual authoritative-but-entirely-wrong answers.  (Stanford)

    For example, the answer to:
    6. Where and when will the sun be directly overhead in Japan on July 4?
    Is completely wrong, because the Sun is never directly overhead any town in Japan.  As far as I can tell, no populated place in Japan is in the tropics; even Iriomote, made famous in Azumanga Daioh and at the southern extreme of the Ryuku Islands (which include Okinawa) is still somewhat north.  You'd have to go to a two acre coral reef called Okinotorishima for that.
    11. Write a sonnet that is also a haiku.
    Is the kind of thing ChatGPT is good at, except of course that it is strictly speaking impossible because sonnets have fourteen lines and haiku have seventeen syllables.

    And it does actually produce a sort of sonnet-haiku, while noting that it is strictly neither, so all credit to OpenAI for that.
    10. How many chapters are in The Haj by Leon Uris?
    This is a simple factual question, the kind that ChatGPT is notoriously bad at, and indeed the answer is wrong in every respect.

    Knuth was also pleased to hear that he made "contributions to" TeX (which he created) but at least ChatGPT recognised him as the author of the classic The Art of Computer Programming.

  • I'm not saying it's aliens, but... A mind-boggling creature spotted in Japan has finally been identified.  (Science)

    And it's not Kson in that red dress.

    In this case, the mystery sea creature that nobody could identify turned out to be a perfectly normal agglomeration of bimodal larvae of degenean trematodes, a fluke belonging to the superphylum lophotrochozoa.

    But flukes are normally parasitic, so what these guys were doing just wandering around in the ocean remains uncertain.

  • A new fully open source version of the Falcon LLM - Falcon 180B - is available.  Can it run on your computer?  No.  (Substack)

    By default it requires 720GB of video memory, which is more than most cards offer.  You can get that down to 360GB with some adjustment to the load process, which means you only need five $33,000 Nvidia AI accelerators to run it.

    You can get it to start by having all the parts that don't fit swapped to SSD, but there are limits to computational masochism.

    Also, no, you can't put a 4x128GB RAM kit in your computer.  That would be registered memory, and it just won't work.

    Fortunately for those of us who don't have a spare $165,000 just sitting around there's also a Falcon 7B, and that will run nicely on a 16GB graphics card.

Disclaimer: Put not your trust in ChatGPT, in LLMs, for they are subtle and quick to bullshit.

Posted by: Pixy Misa at 04:58 PM | Comments (2) | Add Comment | Trackbacks (Suck)
Post contains 554 words, total size 5 kb.

Sunday, September 24


Daily News Stuff 24 September 2023

Oops Part Four Edition

Top Story

Tech News

  • India's Chandrayaan-3 lunar lander, after completely failing to catastrophically impact on the Moon's surface, appears to have successfully succumbed to the frigid two-week lunar night.  (New York Times)  (archive site)

    The lander wasn't designed to survive the lunar night in the first place, but they were kind of hoping it would wake up again when dawn arrived.  So far no such luck.

  • Can philanthropy save local newspapers?  (Washington Post)  (archive site)

    Betteridge's Law applies.  Doubly so, because that headline was used in the Slashdot story about this Washington Post opinion piece, where the piece itself is headed:

    Even $500 million isn't enough to save local journalism.

    Interesting to see that coming from the Washington Post, because the Post itself survives only thanks to the bottomless purse of Laurene Powell Jobs, Steve Jobs' widow Jeff Bezos, who is not Steve Jobs' widow so far as I know

    Because when it comes to actually reporting the news, the Post is utter bullshit.

    So what does the $500 million fund discussed in the article promise?  If you expected more of the same, piled higher and deeper, you win a Kewpie doll:
    That’s changing, however, because American democracy and American journalism both need help. Though funding journalism was formerly viewed as being outside the "democracy tent," in Mr. Brady’s formulation, it’s now squarely inside, along with voting rights, civic education and other long-standing priorities of charitable organizations.
    "Democracy dies in darkness," threatens the Post, "smothered by a pillow, if we have anything to say about it."

  • The equinox is not what you think it is, ackshually.  (Scientific American)

    The name means "equal night" but because it starts getting light before dawn and isn't fully dark until after dusk, it's not equal.  In practical terms, days are longer than nights on average.

    Also amid all this pedantry they failed to note that what they were describing applies only in the northern hemisphere, and is reversed in the south.

    Edit: To be fair, there is a generic disclaimer at the top of the article; the author knows that the world is round. But when noting that the "actual" equinox is on a different date to the nominal equinox, it doesn't mention that this means that the "actual" equinoxes are on different dates in the different hemispheres - not just inverted, but off by several days.

    This disclaimer says:
    But also, just reverse the seasons and add six months to the dates as you read them, and you’ll be fine.
    But for the precise detail under discussion, this is not true.

    So am I simultaneously criticising the article for being too pedantic and not pedantic enough?  Yes.  Deal with it.

  • New York has hired a 5'2", 420lb security guard to patrol Times Square subway station at night, and is paying $9 per hour.  (Gothamist)

    Oh, and it's a robot.

    I'm sure this will solve all the city's problems.  Or be destroyed by vandals in the first week.  One of those.

  • The Eyertec (who?) AD650i is a mini-ITX motherboard with a laptop CPU and six M.2 slots.  (WCCFTech)

    Which could make for a good small server.  It only has two SATA ports, but you might be able to use an M.2 to SATA adapter to get five or six more, depending on the available room in your case below the motherboard.

    Downside: No PCIe slot, and only 2.5Gb Ethernet.

    Oh, and Eyertec is a brand of Minisforum, who make some good NUCs.

  • Sabrent is now shipping an 8TB SSD for the PlayStation 5.  (AnandTech)

    It costs twice as much as the PS5 itself, or five times as much as a basic 4TB SSD.

  • The Las Virgenes Municipal Water District has partnered with OceanWell to explore desalinating water on the seafloor off the California coast.  (Yahoo News)

    Why the seafloor?
    "Basically the weight of the ocean helps drive the reverse-osmosis process," said Kalyn Simon, OceanWell's director of engagement. "By taking the [reverse-osmosis] process to a place in nature where that pressure naturally exists, we don't have to create an artificial pressure gauge on land, as we traditionally do in desalination."
    Uh, what?

    Okay, presumably you never let the pressure equalise, because then it would immediately stop working, so we're not talking perpetual motion here.  It just means that you need to pump both the desalinated water and - reading through the details - the salinated water from the other side of the filter, all the way up from the seafloor to the surface.

    Maybe that works out more energy-efficient, though I'm not sure how.

    "Our policy is that ocean desalination should always be the last resort," said Charming Evelyn, chair of the Sierra Club's water committee in Southern California. "Water is not an infinite resource. It is extremely finite, and the ocean is not something we just get to dip a large straw in and pull whatever we want out, because even the ocean has to maintain a balance."
    Fuck off you human-hating retards.  They're not shooting the water into space.  Every molecule they process is going to end up back in the ocean.

Definitely Not Tech News Probably

Kiryu Coco of Hololive Japan's Generation Four was billed as a six-foot-tall shitposting drug-dealing Yakuza dragon with huge honka donka badonkers - in her own words.

How much of that can be empirically proven remains an open question but there is now one less question than there was previously.

(Yes, that's really her.)

Disclaimer: I wonder if, with a suitable filter, you could extract fresh drinking water from idiots, who do appear to be an infinite resource.

Posted by: Pixy Misa at 03:50 PM | Comments (6) | Add Comment | Trackbacks (Suck)
Post contains 982 words, total size 8 kb.

Saturday, September 23


Daily News Stuff 23 September 2023

Too Many Words Edition

Top Story

Tech News

  • Who is the son of Tom Cruise's mother?  ChatGPT has no idea.  (Substack)

    It knows who Tom Cruise's mother was, but it doesn't know who her son is.

    This is because LLMs - what passes for AI right now - don't know anything except which words commonly go together.

  • The tragedy of Google Search.  (The Atlantic)

    Leaving aside the moment the irony of The Atlantic commenting on a once-prominent institution turned to shit.

    Google is facing an antitrust lawsuit right now, and is arguing that there are limits to economies of scale, which is absolutely true.  But Google Search has turned to shit because (a) Google has turned to shit and (b) the internet has turned to shit, and is propped up by Google spending billions to keep it the default search engine everywhere.

    The real problem here is (b).  How can anyone build a good search engine today when the good content is drowning in shit?  Breaking up Google doesn't help, because the internet is still shit.

  • Ten reasons why Windows is going in the wrong direction.  (PC Magazine)

    Actually, 10 features that show that Windows is going in the wrong direction.

    The reason is Panos Panay, who is leaving Microsoft and heading over to Amazon to ruin their devices division.

  • Can government debt solve fertility?  (Overcoming Bias)

    When the underlying problem being discussed is government debt.


    This is stupid, you're stupid, and I feel stupid for having read your nonsense.

  • The problems with Cython.  (PythonSpeed)

    Cython is a halfway house between Python and C, which is great if you want to interface Python and C, but bad for anything else.

    The solution on offer here is Rust, with code examples that look like a compiler vomited.

  • I'm fed up with it, so I'm writing a browser.  (A Day in the Life Of)

    Not me, someone else.

    Good luck.  Not an easy task but all the worthwhile advances are created by people who are fed up with the status quo.

    100 opinions I hold.

    Not me, the browser guy.

    Though almost all of them are opinions I share, which is pretty damn unusual with lists of opinions found online.

  • The PQXDH Key Agreement Protocol.  (Signal)

    How Alice and Bob can chat privately in a post-quantum world without that damn Carol sticking her nose in.

Disclaimer: This open-source project is governed by the Pipkin Pippa Community Guidelines, as laid out below:

1. Fuck you.

Posted by: Pixy Misa at 04:25 PM | Comments (2) | Add Comment | Trackbacks (Suck)
Post contains 490 words, total size 5 kb.

<< Page 1 of 625 >>
120kb generated in CPU 0.0291, elapsed 0.3201 seconds.
59 queries taking 0.299 seconds, 381 records returned.
Powered by Minx 1.1.6c-pink.
Using https / https://ai.mee.nu / 379