Feed
-
You don't need to believe that AI is intelligent or "superintelligent" to notice that giving it near infinite compute and letting it loose on the internet is quite likely to have bad results, especially for the internet, possibly more widely
-
Does AI have risks? Undoubtedly. Lots of them – from fuelling cyberattacks to military uses to the financial bubble. But to believe specifically in *extinction* risk as the leading threat is clownery, to be honest. "AI takeover" is a sci-fi scenario and "AI kills us all" even more so. There are much more realistic risks right in front of our faces. AND, if we're going to talk about risks to humanity, we should look at the actual science around clima...
-
We can just unplug the computer. Well, hmm, OK. Maybe we'll have to unplug lots of computers. Maybe all the computers. We missed some computers and they turned the other computers back on. We're going to have to destroy all the computers.
-
The two emerging "camps" are both wrong about AI. The extinction people think their sci-fi religion is coming true. The dismissive people think it's all just a hype and a bubble. AI is clearly dangerous. Not existentially so, I think! But still dangerous, most specifically to anything that relies on closely-guarded knowledge/"security through obscurity", which includes most computer systems plus many kinds of real-world weapons. It spreads state-ac...
-
I'm very much a sceptic when it comes to the idea of AI ending the world yadda yadda. But if 'superintelligence' could potentially end humanity, couldn't it also just-as-potentially invent a utopian postcapitalist Star Trek-ish future? Doesn't that possibility deserve at least a little bit of the thought and debate that is being spent on the apocalyptic one?
-
Believing that an AI is conscious or sentient requires you to believe that perfectly comprehensible and replicable mathematics, performed sufficiently quickly, can somehow summon consciousness. The *impression* of consciousness emerges from the input data: the entire web, all obtainable human literature to this point, etc. It's an echo of that vast corpus – nothing more.
-
I set up a DMARC rua=mailto: for a domain, and uh. Why would anyone want this? I'm now getting spammed to heck with reports from 100 different Yahoo domains
-
Minor league baseball is such an awesome feature of America
-
Until I get airtag-like findability for every object in my life I’ll have to stick with the tried and true: • Lose Thing • Look everywhere for Thing in increasingly exasperated and embarrassing fashion • Purchase replacement Thing • Establish home for replacement Thing only to find… original Thing
-
I expected to laugh like an idiot during COYOTE V. ACME I did not expect to have a much deeper emotional response like what the hell. What a picture, great time at the movies y’all.
-
First event post Tim Cook and the new phones come in Alabama Crimson and Carolina Blue that’s stone cold.
-
Last night the preteen asked if we could talk and sat down for an hour of genuine conversation about what’s going on his life that I’m pretty sure was only 10% motivated by wanting to extend bedtime and man. That’s the good stuff right there.
-
RE: https://mastodon.social/@jimray/117138792023999673 Y’all @lexfri is a madman. He sent me a test flight a few weeks after this post and it recently passed my “waiting to take off” test and now it’s a thing. https://lexontech.org/strategery-is-back-and-im-the-developer
-
It’s been at least twenty years since the last time I saw The Roots and tonight they were Hijo Uno’s first concert. Their show’s definitely evolved, still one of the best live bands on the planet.
-
Design Twitter decamping to Threads now arguing about one of their own creating a logo for Meta’s AI like it’s 2012 is a perfect encapsulation of the rot of that industry and also why Threads (and Instagram) are the worst networks because they force you to perform the most obvious self-deception.
-
Lotta middle aged dudes with teenage kids and gray beards at this Gorillaz concert. Wonder what that’s like (to have a beard)
-
Big tech gets all the attention but there are thousands of VC and PE backed companies using AI as a classic arbitrage against bewilderingly complex systems they helped bring into existence to make your everyday experience of the world worse in every way https://www.nytimes.com/2026/09/24/business/ai-hospitals-insurers-health-care-costs.html?unlocked_article_code=1.D1E.Njsb.BTGZ7zu7kzGK&smid=url-share
-
Larry “Friend of Jeffrey Epstein” Summers refused to kick Zuckerberg out of Harvard for the most blatant violation of his classmates' privacy (women, natch) and now we regularly have to deal with credulous news stories from actual journalists about some obvious, awful bullshit he’s foisting on us.
-
The obvious solution is to just ban the gambling apps but I’m pretty sure that’s how we end up with President Dave Portnoy so I propose we tax their revenue at 100% for every dollar over $1,000 they make off a gambler.
-
Angel: Boss we did it, the perfect nutrition for kids. We call it “milk” God: EXCELLENT. NOW MAKE THEM SPILL IT. Spill? Like the parent just… NO. THE CHILD SHALL AT RANDOM SPILL THE MILK. WHILST THE PARENT IS DOING SOMETHING ELSE FOR THE CHILD. Why on earth… I COMMAND IT. jfc WHAT. What?
-
This reads like a pitch for an Apple ad from 1994 that was ultimately passed on for being too maudlin. That it includes a ripoff Miyazaki as the lede art is a perfect fit for how cheap it is. https://x.com/alexandr_wang/status/2103551714536439951
-
It’s not that surprising we’re doing Discourse about mediocre country-inflected pop having a moment after twenty years of dominance by mediocre hip-hop-inflected pop, I suppose I am surprised it’s coming from Rolling Stone not The Daily Caller.
-
I also feel like I am losing what’s left of my mind when smart people who work in media use Facebook’s apps and discourse about it on Twitter as if they have no theory of how media has worked for two decades or agency in the process.
-
One CIO I talked to recently, with 170 people in his organization, measured a 2.1x increase in productivity from AI. That productivity came at a cost. His top users of AI each spent $21,000 on tokens per month. Another engineering leader spends six to eight thousand dollars per engineer per month. If you’re going to make these tradeoffs—and *somebody* is going to—they need to be good ones. Which raises the question: how do you know what you’re getting? https://www.jamesshore.com/v2/blog/2026...
-
As I talk to engineering leaders, I’m seeing *order-of-magnitude* differences in spending on AI. Should you spend thousands per person per month on hands-off agentic coding? Or hundreds on analysis and fancy auto-complete? It depends on the outcomes. You can’t pick the right approach to AI without knowing what you’re getting, and you can’t afford to half-ass this. So how do you measure it? Measure and compare approaches in part 2 of my series on quantifying AI impact: https://www.jamesshore....