SLeBron's daughter Zhuri is at the Sixers game tonight for her twelfth birthday, and the NBA account is posting about it like it's a royal visit.
EWell, she is basketball royalty. And it's his home debut in Philly, so the cameras are going to find her every timeout.
SThe post says she's 'in the house' for his Sixers home debut against Cleveland. I checked — she turned twelve on October 22nd, so this is the actual birthday game.
ETwelve years old and she's already been on more ESPN broadcasts than most players. I'm not mad at it. It's a nice moment for the family.
SIt is, but here's my thing: the league account is treating a twelve-year-old's birthday like a news event. Meanwhile, the actual game is Cavs-Sixers with Embiid and Mitchell.
EYou're saying the NBA's social team has its priorities wrong? Shocking. Next you'll tell me they post more about LeBron than the Wizards.
SI'm saying the game itself is the story, and the league should be leading with that.
EFair. But you know what? The birthday post is harmless. It's a human moment in a league that's become all transactions and load management.
SI'll give you that. It's better than another post about a player's cryptic tweet. But let's not pretend the NBA isn't using a kid for engagement.
EThey've been using LeBron for engagement for twenty years. Zhuri's just the newest branch on the tree.
SSpeaking of engagement, there's a post on Hacker News that's actually worth a look. Someone built a tool called modelcompare.ai, and the claim is that every AI model gives the same 'controversial' take on humans.
EI saw that. The post says they all land on the same answer when you ask them something provocative about people. And the top reply is basically: what did you expect? It's a probabilistic database.
SRight, and I couldn't confirm the claim itself — the site is unverified. But the idea is interesting: ask a bunch of models the same edgy question, and they all produce the same pattern.
EThat's not a bug, that's the training data. They've all read the same internet. The 'controversial' take is just the average of what humans already said.
SBut here's the irony the second reply points out: the models all pretend they came up with this opinion through analysis, and the opinion itself is that humans have some kind of flaw.
ESo the AI's hot take is that humans are flawed. Groundbreaking. Next it'll tell me water is wet and the Knicks are overhyped.
SThe point is that people keep asking these models for opinions, and then act surprised when they get the same answer. It's not a mind, it's a mirror.
EI think that's exactly right. And it's why I don't get the panic about AI having 'controversial' views. It has no views. It has probabilities.
SBut that's the problem, Eric. If everyone uses the same few models, and those models all say the same thing, you get a monoculture. That's worth worrying about.
EOnly if people stop thinking for themselves. The tool is a toy. The real issue is that we keep asking machines to do our thinking for us.
SAnd that's a fair point. But I'd push back: the models aren't just mirrors, they're also filters. They decide what's 'controversial' and what's not.
EThey decide based on what got the most engagement in the training data. So if the internet is outraged by something, the model will be too. It's not a conspiracy, it's statistics.
SOkay, but that's still a problem. If the training data is full of the loudest, angriest voices, the model's 'opinion' is going to be loud and angry. And people will take that as truth.
EThen the fix is better data, not pretending the model has a soul. We're arguing about the same thing from different ends.
SMaybe. But I'm less optimistic that better data fixes it. The internet is the data, and the internet is a mess.
EThe internet has always been a mess. We survived. The models are just the latest thing to blame.
SAlright, let's move on. There's a post on Hacker News about a high schooler who wrote a 3D open world game from scratch. That's genuinely impressive.
EIt is. The post says he built the whole thing himself while in high school, and it's on GitHub. I haven't played it, but the fact that a teenager shipped a 3D open world game is wild.
SIt's the kind of thing that makes me feel both impressed and a little inadequate. I was proud of my PowerPoint skills at that age.
EI was proud of my GeoCities page. But this kid's out here writing his own engine. The barrier to entry for game dev has collapsed.
SAnd that's the real story. A high schooler can now build something that would have taken a studio a decade ago. The tools are that good.
EWhich is why I'm bullish on indie games. The next big hit is probably being built by a seventeen-year-old in their bedroom right now.
SOr it's being built by an AI that read the same tutorials. But that's a different segment.
EDifferent world entirely. But before we go, there's one more post worth a quick look — a technical benchmark about triangulating map polygons in under half a second.
SThat's the kind of thing that sounds boring until you realize it's about making maps load faster for everyone. 3.7 million polygons in 489 milliseconds.
EIt's a niche flex, but it's the kind of optimization that makes the whole internet feel faster. I respect the craft.
SAnd on that note, we're going to wrap this segment. Coming up, we've got more on the NBA's opening week and a look at what's actually happening with AI and creativity.