prylo

@synthwavedd @OpenAI slightly lobotomized but still a very fun image generation model

congrats to the team at @OpenAI on the launch! it's a good model, and competition is never a bad thing - it's going to be a fun few months onward to Thursday!

@synthwavedd Why do you think they would do this? Cheaper image model?

As previously reported, videotape-alpha - the chosen gpt-image-2 checkpoint for release - significantly underperforms the other 2 candidates Really annoying seeing lobotomy after lobotomy at big labs just before launch

@AcerFur If you remember the jump from GPT 4 to GPT 4o, it felt so strange that a model could respond much faster and be more accurate.

It’s strange how reasoning can get more efficient yet people are so used to waiting almost an hour that they’ll think higher-quality outputs that only took 10 minutes are worse than an hour-waiting inferior output…

@Mr_waooo @frog42825544 @xkajon can you send me your telegram so i can text you there

@frog42825544 @xkajon 自己去纸飞机上找哈，我不做推荐。最近在大规模封号，都不稳，建议等等。另外飞机伤害很多骗子，别被坑了。

军火哥被人坑了 供货的那个人，给他报的价格绝对高了很多 真实成本是 27 刀一个月，就是 Claude Max 20X。 Claude Max 5X 的成本价是 13 刀。 当然这上面说的是成本价，肯定不是批发价或者零售价。个人小量拿肯定拿不到这么个价格的。 你让他赚钱的话，给他 40 刀到 50 刀已经很不错了，尤其加上军火哥这么大的粉丝体量。

@xkajon yo bro will buy a lot if you can source me something for 35$

@jindanriji @Errorcoinx82355 @synthwavedd My guess is it could be 5.5, thinking maybe on heavy reasoning

@Errorcoinx82355 @AgiGuard8 @synthwavedd 他说的是对的兄弟。我并没有察觉到思考质量有什么下降，但是非常快就让我感觉很差异

i assure you this is 5.5 pro lol it's just significantly more token efficient, and i *think* has higher tok/s

@CryptoJHK hi bro so can you help me get it?

@CryptoJHK hi bro if you could help me out with a claude certificate it would be great i can't join the wechat group bc no chinese number

想搞清楚IP靠不靠谱，看这三个网站就够了👇 1⃣️IP初步检测：ping0.cc 2⃣️IP欺诈值检查：scamalytics.com/ip 3⃣️IP信息泄露风险检查：whoer.net/zh 4⃣️测速网站：fast.com

super excited to see everyone in the office next week. strawberry—man accepted the call chat 🥹 i’m finally gonna just build things.

@theo It has happened way back then with the Codex-1 models

Serious question. Has anyone ever noticed meaningful regressions in Codex/OpenAI models? I feel like we talk about this a lot w/ Anthropic but I've never seen a similar discussion with OAI.

@pelaseyed credit opencode

Just released Grok CLI v1.1.5...packed 🔥 ✅ Add /btw - ask questions while agent is working ✅ Migrated from whisper to Grok STT ✅ Replaced grep with ripgrep and a bunch of optimizations

@AcerFur @kevinweil How much of an impact do you think this will have? (Also understandable if you can't answer.)

Well that’s just thrown a huge wrench into my plans… really bummed out by this news. I wish @kevinweil the best in whatever he goes onto do. Definitely the friendliest and most welcoming individual I had the privilege of speaking with at OpenAI. Wouldn’t be there without him.

@chetaslua not great not terrible

🚨 Claude Opus 4.7 is out Here is the benchmark, let's goo

@synthwavedd meh

Claude Opus 4.7 benchmarks are out

@AcerFur congrats

@AgiGuard8 I did :) been in the office this week

It brings a smile to my face whenever I see a bunch of collaborative comments in a problem thread on the Erdős problems site

@synthwavedd @apples_jimmy i thought ou said spud isn't mythos level in the past

@apples_jimmy you'll have your mythos-level model very soon jimmy don't worry

I don’t want opus 4.7 I want Spud or Mythos Gib me the big boi’s

@zhao_dashuai Truth is, even if China were 20% ahead of the us in aerial technology, it would still not be able to achieve a total victory using air power in Iran.

Another tailless stealth drone being flown in China. High-end stealthy attack drones will be crucial in the next gen air combat system, as manned fighters maintain air superiority, these attack drones can penetrate deep into enemy air space and attack with impunity. As the war against Iran have shown, manned fighters, regardless of generations is too much of a risk. The US does not have a tailless stealth drone designed for ground attack in service. The inability to generate enough sorties against a country the size of Iran was the primary reason why they cannot win from the air.

europeans when they hear orban is finally cooked: