开启时会模糊预览图，关闭后正常显示
Unbelievably sad bitchy people, on par with Estonians maybe being broken in and then immediately hailed by the West as "civilized" and "advanced" had profoundly confused them, it seems
this moron is hilarious not least because it seems that bamboo is fiercely defended by non-Communist Hong Kongers as their traditional craft (Mainland, I guess, doesn't have a "traditional craft" for residential high-rises) and Mainland hates it. Anything goes to dunk on bugmen
🍌 Nano Banana Pro is very impressive but still seems very hit and miss with resemblance and still have less resemblance than both Flux and Hyper Realism Some photos do not look at all like me (like the first, looks more like Martin Garrix), while others do (the second here) It's also very expensive at $0.15 per photo so it's not super economical yet (Flux for example is $0.01 per photo) For now Hyper Realism on Photo AI still has higher resemblance and is much cheaper, so I'll keep running that. I'll happily switch it over though once Nano Banana Pro becomes good enough or I can hack it to become good enough to have high resemblance (I haven't been able to yet!) Apart from the resemblance, Nano Banana Pro is superior in how real it looks, especially stuff like skin detail, it's really excellent I think in 3 to 6 months though, if there's an update, it probably can have high resemblance and become much cheaper and economical One great thing is that Google deserves lots of respect for is they didn't nerf it at all, it has little restrictions, which is a really great thing, more free expression = better Obviously every technological breakthrough has good and bad, and society will have to adjust to not trust every pic, but it also finally gives us a new time of boundary-less expression which is what I've been working on for the last 2.5 years with Photo AI So it's very very very cool to see it finally happen!!
🇪🇺https://t.co/NdorAWrhrB @euacc 📸https://t.co/lAyoqmT9Hv $125K/m 🏡https://t.co/1oqUgfDEsx $40K/m 🛰https://t.co/ZHSvI2wRou $38K/m 🌍https://t.co/UXK5AFra0o $16K/m 👙https://t.co/RyXpqGvdBB $14K/m 💾https://t.co/M1hEUBB6da $6K/m
Cursor里面的plan模式可以问题式的确认决策，分阶段走 然后执行阶段就交给cc和glm，这个组合用了2周确实好爽，而且用起来一点都不心疼 结果GLM Coding Plan 黑五大促又来了（11.27-12.5） https://t.co/XYiULBfaoX - 包年立减 30%，包季立减 20% - 邀请好友享 40% 赠金 - 年 Lite Plan 只要 168，只要 168，折合一天只要五毛钱！🤣
🚧 building https://t.co/AJfZ3LMlgq https://t.co/606cFUoda3 https://t.co/s0m0tpQMDH https://t.co/UQ5vrrYdAG 🐣learning/earning while helping others ❤️making software, storytelling videos 🔙alibaba @thoughtworks
Do they just think that, now that the REE export controls are suspended for 1 year and so is the BIS 50%, the REE card is off the table and they can just wantonly expand export controls list without technically breaking any agreement?
We're in a race. It's not USA vs China but humans and AGIs vs ape power centralization. @deepseek_ai stan #1, 2023–Deep Time «C’est la guerre.» ®1
one is a post I scheduled in jan 2025 for jan 2026 about “chat with pdf” still being unsolved by then. another is about super long context memory that makes the so appear more sentient to the users by the end of ‘26 because it remembers everything.
making models learn • eXperiments lab • memes and training lores
in jan 2025, I made two leaps of observation about what would happen in 2026. one of them is going to pass with flying colors, the other one is trending towards more likely to fail.
making models learn • eXperiments lab • memes and training lores