Jozi retweeted
Albanian students protest over education costs xitter.cf/i/broadcasts/1rxmqpkXj…
Jozi retweeted
hello 🤗
regretfully i must inform you that multicellular alien life has spontaneously emerged from pure random noise
they keep saying 👁️ ỳ̷̖̹ớ̶̢̫u̴̝͊̊ ̴͇̀̐̐ä̴̰͖̮͌̒ŗ̷͆̀e̶̤͐͆͂ ̸̰̦͑͗̕n̴̹̓o̷͙͊t̶̟̲͈̄͐͘ ̶͂ͅa̶̲͚̾̔̚l̴̝̍̂̑ǫ̶͕͔͆̉ņ̵̪̩̊e̸͖̿ 👁️, but otherwise seem nice enough
1/2 🔊
Jozi retweeted

μck ٤:@JustMicrockOct 3
Europe
EuropeConnected via Europe App StoreAccount-level information, not a live location or per-post device.
im sure this video goes crazy hard if you have room temperature IQ
Jozi retweeted

I Am Devloper@iamdevloperSep 22
United Kingdom!
United KingdomConnected via United Kingdom App StoreLocation may be affected by a proxy or VPN.Account-level information, not a live location or per-post device.
if I was Dumbledore I would've hid the philosopher's stone in the aws console
Jozi retweeted
ngl 90 percent of us burned the Codex Usage Yesterday.
In hope of the Reset and it is almost evening and we still refreshing the page.
Is this a Joke?
Tibo@thsottiauxSep 22United States
United StatesConnected via WebAccount-level information, not a live location or per-post device.
Ladies and gentlemen... start... your... ENGINES. We are almost Tuesday and I promised a reset for Tuesday. Among some other things. See you soon.
Jozi retweeted
Watching people defend Codex burning through paid limits is hilarious.
You paid for the service.
The company is not your friend.
You don’t get bonus points for defending bad value.
Respect your money. 😭
Jozi retweeted

⭕ AI & Design (Marco)@AIandDesignSep 20
United States
United StatesConnected via United States App StoreAccount-level information, not a live location or per-post device.
How long does a Codex $100 plan last? 20 minutes?
This is bad. Already burned 20% of my $200 plan in a day. And haven't used it much. Mostly been coding with Muse. @thsottiaux what is going on?
Jozi retweeted
🚨 Ursula von der Leyen did NOT like that. She just got called out in European Parliament to her face…
"There is no greater RACIST against Europeans than you. You flood us with illegals and suffocate us with taxes!"
Replying to @thsottiaux
Efficient??? Bro touch grass.
I dare you to use only a 200 subscription internally and see how far it gets you.
Jozi retweeted
My TL is now flooded with people complaining that GPT-6 Astra has been “weakened.” 😿
Looking back, when Astra first launched, the TL was filled with amazement and praise.
Now it has almost completely flipped.
It’s honestly both absurd and ironic.
But in a way, it was inevitable.
The higher the expectations were at launch, the harder the disappointment would hit afterward.
Even the slightest difference in experience can get magnified several times over when expectations are this high.
I guess this is what hype backlash looks like. 😹
Even the $200 Pro subscription has now been paused, so a broader GPT-6 Astra rollout in ChatGPT Chat feels even less likely anytime soon.
At first, Astra was astonishing.
“Generational leap.”
“A milestone marking the arrival of the AGI era.”
The praise was enormous, and so were the expectations.
Then, after only a few days, OpenAI quietly began weakening Astra’s intelligence, and the actual experience took a major hit.
I always expected something like this to happen eventually. I just did not expect it to happen this fast.
Worse, some users are being silently routed to weaker models. They pay exactly the same subscription price, yet invisible rules and algorithms effectively turn them into second class users.
Now OpenAI is telling us compute is constrained and demand is overwhelming its infrastructure, to the point that even its highest tier subscription plan has been paused, with a promise to fix things as soon as possible.
So this is where we are now:
> Some users are silently downgraded
> Some users are routed to weaker models
> The highest paid subscription tier has been paused
And even after all of that, they are only barely keeping up with demand.
We went from an astonishing Astra launch to degradation, limits, overload, and paused subscriptions in a matter of days.
Frontier intelligence has never truly been available to everyone.
It wasn’t before.
It still isn’t.
Jozi retweeted
We benchmarked DeepSeek V4.1 Flash by @deepseek_ai .
It reached 98% of GPT-6 Astra’s score at 1.4% of the cost on everyday design tasks based on user requests.
Every model except Astra scored lower AND cost more.
Are open models overtaking closed ones?
Full results below ↘️




















