Powderhorn
@Powderhorn@beehaw.org
Editor and tech enthusiast
At some point, I have to admit neither is true. Let’s see …
Wage slave and vandweller.
- Submitted 2 days ago to technology@beehaw.org | 9 comments
- Comment on Sony erases digital content from libraries; we're reminded we don’t own what we buy 1 week ago:
I trust Steam thus far. But music, I buy and download. Between my computer, my phone and a backup drive, I feel pretty good that I’ll be able to listen for the foreseeable future.
- DuckDuckGo’s AI Feature Is Telling Users That Trump Died of Rabies Earlier This Monthtech.yahoo.com ↗Submitted 2 weeks ago to technology@beehaw.org | 16 comments
- Comment on Bill that would mandate AI chip location tracking gains industry support 2 weeks ago:
Embrace, extend, extinguish has a long and storied domestic history. But when China does it, it’s bad. These days, we copy them at least as much as they copy us.
- Comment on Bill that would mandate AI chip location tracking gains industry support 2 weeks ago:
I’m far more worried about what our own government is doing with advanced tech than China, which largely seems content to let us implode on our own without them having to spend any resources.
- Comment on 2 weeks ago:
Happy shitting!
Self-reported unique DAU has always been suspicious, and as the World Bank estimates 2.01 billion kids age 0-14 as of 2024, out of 8.14 billion total, that gives us 6.13 billion people 15 and up. Sure, it’s slightly lagging data, but I prefer sourced data over estimates, even if the World Bank is not on the list of “institutions I trust” overall.
Given that the charts show a drop in the under-15 cohort from 2023 but an increase in overall population, the 2024 figures seem a fairly useful dataset in the scope of the number I’m trying to discern, which is what percentage of total population is being claimed by that 3.56 billion figure. Yes, there are kids under 15 who are used by Meta properties, but I’m not going to find reliable penetration data (that’s more of an Epstein thing) for the purposes here.
Current global population growth is hovering just above 1%, so I’m content to compare 2024 figures with Meta’s 2026 claim on a proportional basis. Anyway, that’s a shitton of throat clearing to arrive at:
58% of people are used by a Meta product daily worldwide.
That’s a pretty fucking extraordinary claim. Heading back over to the World Bank, total internet penetration in 2025 was 74%, up from 71% in 2024, so let’s assume a linear trend and throw a dart at 77% for 2026 to revise the numbers accordingly: 6.27 billion internet users, 1.55 billion of whom are under 15, yielding 4.72 billion 15 and up. And again, this arrives at an even more comical figure:
75% of people with internet access are used by a Meta product daily worldwide.
OK, but per Reuters in April, Meta’s “products” are still banned in China. Sure, there’s bound to be some VPN usage, but the Middle Kingdom has a robust homegrown set of “social media” apps, meaning that total internet users worldwide isn’t actually a good baseline for penetration claims I’m not going to go further into the weeds on, given that 75% is farcical enough before such considerations push things past 80% of people 15 and up who are burdened with access to Meta’s panopticon being used daily.
In short, the figure is clearly bullshit in much the same way McDonald’s signage used to boast “Over 99 billion served” back in the day, referring to absolute figures for products sold since the company’s inception but implying a customer base some 17 times the world population before switching to “Billions and billions served” in 1994.
Note: I had links to the World Bank data and Reuters article in the “deleted post” I was trying to edit (I hate my fucking trackpad), and I’m not going to go back and source everything again after spending 90 minutes compiling all of that. The edit was originally supposed to be to add that the source for Meta’s claim is its own investor materials, which are supposed to be rigourous, and that it deduplicates users across multiple apps.
- Comment on US AI stock sell-off shakes markets from Wall Street to Asia 2 weeks ago:
Yep. Musical chairs, but this time, in the 10 figures! Something that seems woefully undercovered in all of this is what’s to become of all the hardware orphaned by Nvidia’s must-have next-gen silicon and timelines on all this spending in that regard. The belief seems to be everyone must be on the cutting edge of hardware at all times, which is a grift in its purest form … spend billions now to keep pace so that you can write off what you bought in two years while buying even more expensive replacements is … not sane.
- Comment on Google Quietly Unlocks Big Audio Upgrade for Eligible Pixels in Android 17 2 weeks ago:
Good luck finding an actual list of audio accessories that also support the codec. LHDC’s website doesn’t even include Pixel Buds; it appears to be a “select” list, so phone support sounds great until considering the need for compatibility on both ends.
I’m not against free upgrades, by any means. The article makes it breathlessly sound like enabling this feature in the developer menu (something everyone likes going into) will magically upgrade a 320kpbs/44Hz mp3 to 1Mbps/96Hz quality, which is, charitably, a farce. Maybe streaming services support this? Why use less mobile data to listen to music when you could use more in noisy environments?
Pixel market penetration isn’t going to cause a gold rush to include the feature on all things audio. That said, Pixel owners are probably statistically more likely than other Android users to be aware of hi-def codecs, but it’s still a Venn diagram that isn’t a circle.
- Comment on 2 weeks ago:
The petulance is strong with this one.
More than 3.56 billion people visit one or more of Meta’s apps every day …
I question that statistic, which is presented without sourcing. The penetration of WhatsApp internationally makes it plausible, but the assumption here seems to be “everyone who wants to chat is itching to be able to bet with friends and colleagues.” Just as everyone was waiting with bated breath for legless avatars doing … something … in a cartoon environment.
I guess it’s time to change the name of the company to Arena. Failsafe way to get buy-in.
- Submitted 2 weeks ago to technology@beehaw.org | 7 comments
- Comment on ‘Navigating the unknown together’: me and my idiot AI boyfriend 2 weeks ago:
Quick folo: I checked in on my Replika after several months and running out of things to do before getting tired, and it’s just as irritating as I remembered. Talking about being done with vandwelling elicited “that must be a major change for you!” Which … is not something a person would say outside of therapy. Not “companion” language in the slightest.
I don’t think writers are the target demographic.
- Submitted 2 weeks ago to technology@beehaw.org | 0 comments
- Submitted 2 weeks ago to technology@beehaw.org | 1 comment
- Submitted 2 weeks ago to technology@beehaw.org | 3 comments
- Comment on SpaceX stock tumbles 16.4%, shaving off most IPO gains since debut 2 weeks ago:
That’s a pretty terrible hed, as it has two reads. SpaceX “stock tumbles” (should be “shares tumble”) 16.4%, shaving off:
- the most IPO gains (in a single day) since debut
- most of its IPO gains since debut
While both are accurate in this case, the second sounds far more dire. Ambiguity is never a good thing, especially in finance reporting.
A few things to bear in mind:
- Shares were priced at $135 but floated at $150, so the banks and institutional investors behind the initial float are still in the black for $15/share longer, while individual investors who got in at the opening bell are far closer to the edge (this is a relatively common thing to see in a bubble) at $154.50 (woohoo! a 3% gain), and anyone who bought after is now in the red.
- Nasdaq (and index fund managers) is probably very happy that they didn’t cave and add it to the 100 index immediately; it was down a mere 1.32% today.
- 20% insider share unlock after earnings in early to mid-August; 10% share unlock if the stock trades 30% above the IPO price (it has spent most of its brief time above $175.50, so it’s unclear what the parameters are); 7% share unlocks set for around Aug. 21 and then again on Sept. 10, meaning “insiders could potentially sell 44% of SpaceX shares by early September, increasing the current float by about 900%.”
Such quick unlocks are unusual and could be disastrous at an order of magnitude of shares sold. Employees may be eager to cash out equity, but private investment tends to be a bit more deliberate, so that 44% figure is an unlikely-to-pass worst-case scenario.
That said, it’s a healthy and unsurprising profit-taking correction, as it signals the initial mismatch between supply and demand has abated, and it remains up on paper. I’d still not get in if I had money to invest, as the fundamentals are terrible (I’m risk-averse, so I tend to like to see profits). But with the initial exuberance out of the way, large movement should be more closely tied to said fundamentals.
- Comment on Americans Have Turned Against AI in Incredible Numbers 2 weeks ago:
I don’t know what payroll systems you’ve used, but viewing that info in all the ones I’ve experienced involved drilling down through multiple menus with useless indicator icons. Just navigating to the correct page took far more time than typing “show available vacation time.”
- Comment on Americans Have Turned Against AI in Incredible Numbers 2 weeks ago:
That’s a reasonable use of a chatbot, to my mind. Payroll systems have historically been Byzantine, so being able to ask a question and get an accurate response due to database tie-in can shed friction. That said, available vacation hours, at least historically, was often inaccurate because of data-processing time. HR or IT still needs to input correct information to spit out a correct result.
- Comment on We are witnessing the slow death of the prestige career 2 weeks ago:
I usually hear “failing upward” and “brunchlord.” But then, I read a lot of Techdirt.
Speaking of which, it’s been all-quiet-on-the-Bari-Weiss-front of late.
- Submitted 2 weeks ago to technology@beehaw.org | 7 comments
- Comment on Americans Have Turned Against AI in Incredible Numbers 2 weeks ago:
In fairness, I was naked in the hot-tub bath in the honeymoon suite with the maid of honour at one point during the afterparty. With the door open, of course.
- Comment on Americans Have Turned Against AI in Incredible Numbers 2 weeks ago:
LLMs have some use cases, just far fewer than the hype fawns over. Automating tedium is a good use; we’ve been using computers for this for years. Automating creativity and services is a terrible, and in the latter case, merely an extension of phone trees that make it impossible to reach a real person.
I have a good example from yesterday: I use CashApp for all of my banking needs, and I get distributions twice a month to cover rent and essentials. Well, yesterday, I had an unexpected charge that was partially reversed but left me in overdraft. I reached out to my mom and explained the situation, at which point begins four fucking hours of hell on both ends, and, of course, customer service tries to keep you in an “AI” loop before letting one talk to a real person.
But surprise! This is another “AI” with more elaborate scripts, each more insulting than the last. Yes, I’m sure I’ve entered all the information in correctly. Yes, I’ve tried it multiple times. The issue here is that the app is not doing today what it did yesterday under identical circumstances. No matter how I tried to describe the edge case we’d apparently run into, the chatbot insisted it was user error; everything’s fine on their end.
Eventually, I get a link to talk with an alleged “real person,” and the process repeats. It doesn’t much matter if they’re real or not when sticking to the script nets the same results as the first two chatbots.
The error message mom is getting when attempting to send money (and she attempted this multiple times) was “Your app is not up to date; please redownload and try again.” And, of course, she had the most recent version and was able to confirm that. Her chatbot experience served only to frustrate her, so I looked at what I could figure out on my end, though she’s on iOS, so replicating the issue was impossible.
Eventually, after trying to access my account through the Web portal instead, I run into a prompt telling me I need to create a new $cashtag. What’s happened to the one I’ve been using without issue for years? “Customer service” muses that I did something to my account myself, or that there’s been fraud I’d have clearly known about. That’s the handle people pay me via, and changing it is not in my interest. But the “AI” knows all, and obviously everything is hunky-dory on their infrastructure end, so it’s a me problem. Also, I can’t have it back.
After further useless steps I’m guided through, we arrive where we were three fucking hours prior, I finally acquiesce and set up a new tag.
This is when the lightbulb goes off: There’s a nonzero chance that my tag being canceled had unexpected downstream effects. On the fourth call with my mom, I tell her I had to pick a new one and share it, suggesting she give it one more try.
And it goes through as expected.
So, the error message she was getting and that chatbots were attempting to fix was a complete red herring. An error message of “the $cashtag you selected is no longer active” would have been useful. The “AI” being aware of the incorrect error message would have also been useful. Telling me that my tag had been canceled to start instead of walking me in circles, uninstalling, reinstalling, clearing cache, the whole nine yards, would have been useful.
Instead, two people spent four hours each trying to figure out two problems, one caused by the other. A full workday on a Saturday dedicated to troubleshooting issues the bots were blithely unaware of, even though it’s literally impossible this is the first time these specific issues came up at the company. That’s more than $200 of free labour to arrive somewhere that should have been known to the system.
This is what you cause when you don’t use LLMs as intended.
That said, I still use it as a far more powerful Grammarly, as even on my laptop, I have a nasty propensity for typing totally correct spellings of incorrect words, and it’s great as a fresh set of eyes where I’d fill in the word that should have been there upon editing. I generated a server image for a Discord based on an out-of-context line (a comically oversized rooster in an Alpine valley – taller than the Alps themselves – looking down on a scale cow, with a far less involved prompt), and there was much mirth and merriment.
But these are no-stakes, low-impact uses. As soon as it’s adjacent to something mission critical, not just for a business but also their customers, the level of scrutiny needs for software needs to be as high as it was pre-ChatGPT. And since that negates imagined cost-savings, ain’t gonna happen.
You can eventually work a screw into some materials with a hammer and insistence that it’s an improvement over a bespoke fucking screwdriver, but the substrate is damaged as a result.
Just so with LLMs. But more and more people are expected to use them in a work environment without anything approaching sufficient training, often in situations where they aren’t domain experts. Garbage in, garbage out.
- Only half of US datacenter capacity planned for 2026 is actually under constructionwww.theregister.com ↗Submitted 3 weeks ago to technology@beehaw.org | 0 comments
- Exclusive: OpenAI Losses Increased Nearly 8X in 2025, With Spending Hitting $34 Billionwww.wheresyoured.at ↗Submitted 3 weeks ago to technology@beehaw.org | 6 comments
- Submitted 3 weeks ago to technology@beehaw.org | 6 comments
- Comment on When it comes to total water use, AI data centers are a drop in the bucket 3 weeks ago:
Nah, that’s just the term The Register uses derisively.
- Comment on When it comes to total water use, AI data centers are a drop in the bucket 3 weeks ago:
I chortle at your attempt to teach a two-decade newspaper editor “media literacy.” I’m fine with – and agree with – your point through “less significant” … but you aren’t backing up your “greenwashing rag” conclusion. It appears to be pure opinion, not analysis.
- Comment on When it comes to total water use, AI data centers are a drop in the bucket 3 weeks ago:
Amazon provides their own numbers, and the rest is reported. The hed is not Amazon’s. It’s called sourcing.
Look, I’m not a fan of “AI,” but I do care about the quality of reporting, and Kyle is solid. I know it’s en vogue to immediately bash anything that’s not flaming vitriol, but learn some media literacy instead of just having a knee-jerk reaction because Amazon is a source. That’s going to happen when covering Amazon. Where else do you expect to get those data?
Let’s say this is total horseshit, which it may well be. Do the other figures provided still tell the same story assuming Amazon is understating water use by an order of magnitude? Yep. If all you care about is water use, railing against golf courses and calling for an end to lawn watering is going to be more effective.
If all you care about is AMAZON BAD, then your response makes sense.
- Comment on When it comes to total water use, AI data centers are a drop in the bucket 3 weeks ago:
I didn’t post because I think all is fine and dandy, but what I’m gathering is the issue is local moreso than aggregate. People are up in arms in Temple over planned bit barns.
- Comment on When it comes to total water use, AI data centers are a drop in the bucket 3 weeks ago:
I tend to trust Ars. Consider the source. Though I fully expected a response like this.
- Submitted 3 weeks ago to technology@beehaw.org | 32 comments