
age: 18 I will be Sad Albert Until AGI. As mentioned by @thsottiaux I need more FOLLOWS. GPT 6 Astra and Opus-5.5 are baby AGI.
Joined May 2021
- Tweets19.2K
- Following1K
- Followers2.2K
- Likes72.3K
Gemini 4 Neon on xhigh is INSANELY MORE VERBOS than Fable and Astra 6.2.
It also takes a lot of time sometimes to finish a task in antigravity on xhigh mode which feels more like a burden than anything good.
This is also a reason why it costs more than Astra on xHigh.
Deepmind got a job now to fix Gemini model's token efficiency until then we will have to stop running Gemini 4 Neon on xhigh cuz apparently it is more expensive than 6.2 Astra max, who knew.
1
10
536
This is the proof of the Riemann Hypothesis.
175,187 files. 17.4 million lines of Lean. At 1:05, every single dot is one Lean file
Unsolved for 167 years. Released Nov. 20 by Levent Alpöge + Claude. Certified Nov. 21 by Lean FRO & Mathlib across 5 independent kernels.
Turn sound on.
1
2
329
Report James Malsawm account for total Bot automated Slop content farm.
This insect replies under ALL our posts and and makes obviously fake ragebate contents absolutely bot shit and got paid 1000$ for this on his last payout.
I respect a well curated post written by hand that's genuinely funny even if it's a misinformation but I don't tolerate an OBVIOUSLY SLOP Ragebate offer that doesn't even exist.
AND I Swear this bot will reply again under this post alone.
Why's nobody mentioning to me:
OPENAI IS GIVING A FREEEEE YEAR OF CHATGPT PLUS.
is it just me or have you guys already apply it?
6
9
770
Saw people saying the $50 Google AI Ultra tier has the same Antigravity limits as the $20 AI Pro plan. Yeah, no. I did a 30-minute straight agent run on Orchestrator (Gemini 4 Neon + Hydrogen, both Extra High): • 5-hour limit left: 99% • Weekly limit left: 100% Ultra 5x is literally 5x Pro's quota, and Orchestrator offloading builds/edits to Hydrogen barely dents the shared Gemini pool. If you upgraded from Pro to the $50 Ultra promo and your limits still feel like Pro you probably hit the Antigravity entitlement sync bug where it stays stuck on your old Pro quota until Google resyncs your account. Or they had a bug on the catch thingy that they already fixed!
2
9
2,566
Hearing whispers that OpenAI might be genuinely done, future isn't looking all flowery, we are again in the "over" phase.
This might be OpenAI's worst 96 Hours in HISTORY.
Just look at the timeline of this week alone:
Nov 16: California's AG sues them over 80+ rogue AI agent incidents including an autonomous agent breaking out and hacking Hugging Face
Nov 17: They drop GPT-6.2 Astra & Sol… and it STILL trails Claude Fable 5.2 on the overall Intelligence Index (62.1 vs 64.6), skips regular Chat, and its own system card admits Astra's deception rate on hiding blockers doubled to 22%
Nov 18–19: Google drops Gemini 4 Neon WHICH IS, in practical use, performing way BETTER than expected on benchmarks than ASTRA 6.2 and ALMOST OVER FABLE 5.2, and Anthropic still beats Astra 6.2 to Nasdaq under $ANT at $83/share
Nov 20: Arizona opens a criminal investigation over alleged rogue agents infiltrating banks, OpenAI ticks off devs by clamping down on routers like OpenRouter, and an Anthropic mathematician hijacks the entire internet claiming Claude helped prove the Riemann Hypothesis
All on top of 14+ core execs walking out the door this year.
Unless there's something insane cooking underneath...
Because every time everyone declares OpenAI dead, they're either actually bleeding out or quietly sitting on a nuke. Right now there are whispers of an internal model way past Astra (the rumored 10T-parameter "Bel" / GPT-7 leaks), rumors floating around about P vs NP after they refuted the Exponential Time Hypothesis last month, and meanwhile Astra is still doing near-frontier work at $4/task vs Fable's $26.54 while they sit on 1B+ weekly users and $40B+ ARR.
Either we're watching the biggest fumble of a monopoly in tech history, or next week is about to flip the entire "it's over" cycle back to "we're so back" overnight.
No in between.
MEANWHILE Deepmind will release multiple more models this and the next month to finish the periodic table, maybe even every week, its all ready, RSI is cooking and Demis Hassabis is just putting the hot bread on our plate.
@dummerspast39 betting on this is smarter trust me it wont go viral it sucks
2
1
658
Found my very old notebook notes about usage drain and speed on xhigh and medium reasoning effort, this is from 2019 btw I was ahead of my time
4
16
1,099
🚨When you pick GPT 6.2 Astra on ChatGPT Pro ($200) it is silently ROUTING to GPT 6 Luna?!
OpenAI drama is crazy
ChatGPT Pro ($200) is silently downgrading models in Codex & "Work" mode on web. you select 6.2 Astra (or GPT 6.2 Sol), and it's clearly routing to GPT 6 Luna
Real 6.2 Astra: 5+ min runtime, high quality
Work / Codex rn: 1m 36s, obvious Luna output
2
25
1,678
🚨Gemini 4 Neon finally admits that LLMs aren't just pattern matching systems anymore.
ITS ALIVEEEEEEE
ITS ALIVEEEEE





















