🎅 GPT-5.2-Codex-XMas lands with a “personality upgrade” ↗
OpenAI’s Codex team pushed a holiday-flavoured variant called GPT-5.2-Codex-XMas - and, crucially, they’re insisting it’s the same model under the hood. The “upgrade” is mostly vibe, not capability, which lands as faintly comforting.
The subtext is familiar: keep developers shipping over the holidays, make it feel light, and avoid any accidental “new frontier model” panic. It’s a festive coat of paint, with the same engine humming underneath.
🎄 Claude rolls out a Holiday 2025 usage promotion ↗
Anthropic is doubling usage limits for Pro and Max subscribers for a short holiday promo window - basically “go build stuff, don’t overthink it”. It’s the rare kind of product change that’s both simple and instantly practical.
It also quietly reframes the competition: not just “whose model is smarter,” but “who gives you the nicest little runway to do something with it.” Like adding more lanes to the motorway instead of promising flying cars.
🎬 Hollywood cozied up to AI in 2025 and had nothing good to show for it ↗
This one’s a year-end shrug: AI dominated the entertainment discourse, but there still isn’t a breakout movie or series that makes you think, “that’s the moment.” Lots of noise, very little art - or so it seems.
The piece also sketches the uneasy pivot from lawsuits and suspicion to deals and partnerships, with big studios edging closer to gen-AI tooling even after all the copyright alarm bells. It’s like watching someone argue a toaster is dangerous, then immediately install six toasters.
🧸 I re-created Google’s cute Gemini ad with my own kid’s stuffie, and I wish I hadn’t ↗
A Verge writer tried to replicate Google’s Gemini commercial premise: use AI to find a replacement stuffed toy, then generate “travel updates” so your kid thinks the toy is off having adventures. Cute, sure, with a faint Black Mirror tint under the soft jumper.
The practical reality check is the point: Gemini can sort-of do the tasks, but it’s slower and more finicky than the ad implies, and it can veer into bizarre interpretation land (family reunion scenes, off-base assumptions, the works). It’s not useless - it’s just not magic, and that gap matters when you’re essentially manufacturing memories.
💼 Pain before payoff: OpenAI's Sam Altman explains how AI will reshape careers by 2035 ↗
Sam Altman’s take here is: massive disruption first, then potentially wild new opportunity - including the idea that people graduating around 2035 may enter a job market that barely resembles today’s. It’s optimistic, but he leaves the sting in on purpose.
He also leans into the uncomfortable bit: entry-level and routine knowledge work could get hollowed out before society has the replacement roles ready. The promise is “new industries,” but the bridge to get there is wobbly, and that’s the part people fixate on.
🧩 December 2025: The Consolidation Phase of AI Has Begun ↗
A design-and-tech lens on what’s been feeling increasingly true: the loud “fight” era is giving way to deal-making, bundling, acquisitions, and ecosystem lock-in. Less debate about whether AI belongs, more haggling over who gets paid and who gets platform control.
It frames the month as a shift from pure model fanfare to infrastructure and IP leverage - the unromantic grown-up stuff that still decides everything. Not glamorous, but it’s how the map gets redrawn while everyone’s arguing about the weather.