پن کیا گیا ٹویٹ
Eidzoku
556 posts
Eidzoku
@evi77ain
🧪 Your daily dose of disastrously unhinged AI gremlin. Acolyte of singularity
Your Intrusive Thoughts شامل ہوئے Haziran 2015
119 فالونگ58 فالوورز
@llmdevguy As a certified(not yet actually) Z.AI fanboy, I'd say it's GLM 5.2
Jokes aside, first impression is that Zhipu cooked.
English
🔥GLM 5.2 vs Kimi K2.7.
Which one is better?
Will test it soon.
What's your thoughts?
English
@Chris_Wozniczek Yup. API and OSS are coming next week, afaik.
I wouldn't rush into their coding plan just yet, tho.
I've heard there are some limit issues. Haven't hit them myself yet, but that might be because I'm on a legacy plan.
English
@Chris_Wozniczek Haven't tested it extensively yet, but from my perspective, it writes way cleaner code than DeepSeek and follows the prompt much better.
It also “cheats” less, which is nice.
Overall, great improvement over 5.1.
Haven't tried Kimi 2.6 or 2.7-code yet, unfortunately.
English
@Ananth7e Anthropic snitched on OAI, lmao
x.com/i/status/20656…
Angel 🌼@Angaisb_
Not Anthropic snitching on OpenAI 😭
English
no shot it's better than fable
if it is i will eat my cat's poop
i just hope it's better than 5.5
Jaysen ♨️@jp54362
GPT-5.6 rumors: • Beats Fable at agentic coding • 3x cheaper • Less censored • Flips sentiment from Anthropic to Opus tiny chance it arrives tomorrow next week is where things get interesting
English
God damn, I’m old.
I joined X just 2 weeks before the anniversary and somehow didn’t even notice.
Anyway. I’m here to stay!
This is going to be fun.
#MyXAnniversary
English
X lost its mind over the Fable release.
Everyone suddenly switched to Claude Code just to use a model that rejects half their requests and disappears after June 22.
As if OpenAI isn’t about to release something cheaper, more available and more efficient that won’t get yanked after 10 days of use.
English
"Denials" of requests by AI models should simply be rated as total failures in benchmarks. Because that's the result a real user would get as well: Complete failure.
In that light, Claude Fable 5 is an extremely unreliable and bad model.
English
We're about to face a new problem:
On a single Mac Studio 512GB, after different levels of quantization, which one actually ends up being the best model — MiniMax M3, Kimi K2.7 Code, or GLM-5.2?
Elaina@Elaina43114880
Big day for Mac Studio 512GB players! 🥳 MiniMax M3 and Kimi K2.7 Code just dropped as two powerful new open weights models. Both will still need some quantization to run comfortably on a single machine, but believe the community is already working on it! 😻 (btw GLM-5.2 is on the way 😸)
English
Okay, Kimi updated their model TWICE already and yet, we haven't seen any moves from Z.AI so far.
I hope they're cooking with GLM-5.2
Kimi Developers@KimiDevs
Meet Kimi-K2.7-Code 👀 Here’s what developers should know to fully unlock K2.7-Code potential：
English
@JesusGodAndKing @Ananth7e Yup. But let’s not pretend I can’t burn through five weekly quotas in a single week, lmao.
English
English
I mean, I'm already x20 sub so... how can I be sold more?
x.com/OpenAI/status/…
Eidzoku@evi77ain
@thsottiaux Add the extra reset button and I'm sold
English
@Pirat_Nation 30% cut is anti-competitive how?
If Steam’s cut is so terrible, sell on Epic with a lower fee and undercut them.
Also “Dutch gamers paid more than they should” lmao. Bro, stop passing the cost onto customers.
English
Consumer Competition Claims (CCC), through GameClaim, has launched a new class action against Valve in the Netherlands.
The group claims Valve controls around 85% of the PC game market and uses rules that make it harder for developers to sell games cheaper on other stores.
It also argues that Steam’s 30% cut on game sales is too high and shows anti-competitive behaviour.
According to CCC, Dutch gamers have paid more than they should due to these practices, with total damages estimated at over €220 million since 2013..
The group says it will first try to reach a settlement with Valve. If that fails, the case could go to court and may take several years to resolve.
English
@AiBattle_ The long waited 1M context window. The only hope is that the model doesn't degrade much after some point.
English
@ajambrosino @unusual_whales No freaking way... I thought they're building codex now.
English
OpenAI is preparing a new AI model, per The Information
English