And what makes me the most irrationally angry is that Don't Look Up has the single best joke about academia ever with Jonah Hill's character: "Where did you guys say you're from"
Leo: "Michigan State"
Jonah Hill: "Can we get some folks from Harvard to weigh in"
Leo: "It's a good astronomy program.."
I can't stand Don't Look Up after enjoying a huge chunk of Adam McKay's career. The transition to smug seriousness and "look how much smarter I am than everyone else" after movies like Talladega Nights and Step Brothers is weird and I don't have any taste for doomerism in this political era
me: beaver
Scientist: not a fish
Pope: ACKSHUALLY
Golden age of platforming with other games like Banjo Kazooie and Banjo Tooie, not to mention Bomberman 64, Kirby 64 the Crystal Shards, Conker's Bad Fur Day and all the ones you listed (also Majora's Mask remains one of my favorite games of all time so completely agree with that inclusion)
okay close the channel there’s never gonna be a better one than this
i heard the story from a guy who used to lead a flavor research department
scientists developed a flavor compound that was ~fruity but didn't taste like any existing fruit
to help sell it, the idea of "blue raspberry" was invented. didn't matter that that's not a real thing. people liked it.
It seems a curling scandal is sweeping the Games.
I'm Director of IT at my company & I have been pushing us into going paperless wherever possible, I've mostly sold it as a money saving & green initiative, but the primary reason is I want to minimize the amount of time I or my team spend dealing with these infernal machines.
ok this is pretty good
I like this rule of thumb for development with LLMs. I think there are things where I personally need to broaden this slightly (I know there is a problem in this Makefile, I know it's happening here, but I can't figure out the right Makefile syntax so let's ask Claude),but it's a good theory of mind
❤️
I think "lowering the activation energy" is a very good phrase to describe this feeling (I say as a physicist lol). I have overcome writing blocks just typing in single sentences and saying "I am thinking of a word to describe X thing, but it's not Y" instead of spending minutes with a thesaurus
"somehow, Palpatine returned" the most insightful work of civic commentary of the decade
I gotta write. But how write when doomscroll never stop
A line graph titled "Emotions While Coding With Claude" plotting enthusiasm over a 15-minute timeline. The graph features a purple line that oscillates wildly and chaotically between two extreme states labeled on the Y-axis: the peak is labeled "Programming will change forever," and the trough is labeled "This is the stupidest machine I've ever interacted with." The visual illustrates a rapid, repetitive cycle of extreme optimism and frustration.
a photo of Lu Xun
Apropos of nothing, remembering the old Chinese author Lu Xun:
“Lies written in ink cannot disguise facts written in blood.”
― Lu Xun
Anthropic made a very conscious decision to not do any kind of image or video generation unlike OpenAI and XAI, which 1. cuts off a potentially really ugly legal liability tail, 2. is much cheaper to train and serve, especially without video, and 3. makes the product far more enterprise-friendly.
There should be a rule that if you are a Known Sports Poster you can say things like “holy fucking shit” with zero context but if you are primarily a Known Politics Poster that gets you like a two day ban
It's always such a joy to find out people I follow for unrelated brilliance are also fans of Critical Role
Literal Immaculate Reception
he integration of Gemini into their larger toolchain seems to be hurting the perception amongst my colleagues and friends. Even my friends doing very significant scientific software development and getting benefit from tools like Claude hate Gemini integration in Gmail.
Will Stancil after he time-travels back to spend his life with Peggy Carter
I had my computational astrophysics attach chat logs if they used AI and I have to say I got tremendous joy from a couple of them where at one point a student told GPT "this seems more complicated than what my professor was doing" and then GPT course corrected to giving more helpful advice
Honestly? Congratulations, your hardware *and* software are both working incredibly well and responding nearly instantly. All QA testing is a kick in the balls. (Tellingly, many people’s reaction to this is schadenfreude about AI, though this robotics demo likely has little to do with LLMs.)
I‘d be happier if I ran the world and the people who run the world seem like they’d be happier if they were mildly popular on social media, so perhaps we could work out some form of trade.
Sometimes there's even a helicopter involved!
This is a nice write up. It's painful to read and continue to be reminded that he's gone, but the NYT talking to Steve and Alex and Ellen makes me feel like Nuno's life is being eulogized in a way that it deserves, commensurate with the profound positive impact he had www.nytimes.com/2025/12/20/u...
