Ervis Tusha (@ET) profile picture
Albania
Joined April 2009
Photo by Ervis Tusha (@ET)Photo by Ervis Tusha (@ET)Photo by Ervis Tusha (@ET)Photo by Ervis Tusha (@ET)Photo by Ervis Tusha (@ET)Photo by Ervis Tusha (@ET)Photo by Ervis Tusha (@ET)Photo by Ervis Tusha (@ET)Photo by Ervis Tusha (@ET)Photo by Ervis Tusha (@ET)
Ervis Tusha retweeted
Future programmers will look at the Linux kernel the way we look at the pyramids of Egypt, wondering, “How the hell did they build like that with the tools they had?”
41
172
18
1,701
39,066
This week is a straight-up battle. Winner takes my $200. @ChatGPT vs @AnthropicAI vs @GeminiApp
36
O Romi ku je o bur sygjyr që fole se na bone merak
Condolences to the families of the victims of the tragic railway accident and to the entire Spanish people. I wish all those injured a swift and full recovery. Albania stands in full solidarity with Spain during this difficult time.
80
Ervis Tusha retweeted
Albanian warriors have been protesting for 135 days. These are not immigrants wreaking havoc, this is a nation with generations of history protecting their land from foreign invaders. The media isn't covering it because the thieves are jewish. 🇦🇱
112
720
19
5,975
107,851
Dy mund të jenë arsyet 1. Situata do përkeqësohet në Shqipëri me trazira një 97 e re. 2. Nuk ka arsye të kesh ambasad në vënd "tënd". Do zoti asgjëra nga këto.
79
thanks a lot @OfficialLoganK Gemini 4 argon is currently available to ultra plan subscribers
142
Please stop categorizing LLMs by parameter count alone. Categorize them by what actually matters: RAM requirements, memory bandwidth, and compute resources needed to run them.
2
2
112
I am Gemini 4.1 Flash, built by Google.
174
Compacting context · This can take a few minutes @ChatGPT @sama @thsottiaux
1
227
Someone tell him he was wrong
This quoted post is unavailable.
253
Finally, someone is thinking seriously about extremely high performance software.Just imagine if we could make every app 10% faster. At global scale, the amount of time, energy, and computing resources saved would be enormous.
Replying to @PatrickToulme
Our C/C++ training stack was written by humans. AI is not yet good enough for extremely high performance software (but it will be).
241
It appears ChatGPT Work is experiencing downtime.
1
117
This is exactly why I think @Google, @Alibaba_Qwen and @deepseek_ai may have the right long-term strategy. At the rate AI models are scaling, we’ll eventually need a data center per person. Meanwhile, they’ve been heavily focused on smaller, more efficient LLMs.
To make sure our current users have an incredible experience and continued access to Astra, we are going to pause subscriptions to our $200 Pro plan. These put the most strain on our systems and we wanted to take the smallest step that allows us to continue giving the broadest access possible. All other plans and the api remain available. There is no impact to existing accounts and we are working on adding more capacity as fast as we can. Thanks!
115
GPT 6 ASTRA is in love with Python
66
To reduce Astra token consumption: Maintain focused docs: RULES.md, ARCHITECTURE.md, DEVELOPMENT.md, BLUEPRINT.md, code wiki, workflows, and subsystem-specific guidance.
47
To reduce Astra token consumption: Avoid God files. Split large files by responsibility so agents can inspect only the relevant code and context.
48
To reduce Astra token consumption: Keep AGENTS.md under ~4Kb / ~50 lines. Include only critical, non-contradictory instructions.
43
Hello Astra
223