Pinned Tweet
Vivek
46.3K posts
Vivek
@vivekdevops
🚀🛰 🇮🇳 தமிழன் - கொங்குநாடு - 🏹 Tiruppur 🔄 Bengaluru | #WootzSteel #kongu #TVK $ZEC #ZCash https://t.co/3sT0X0flcb
Tiruppur Joined Haziran 2010
7.5K Following1K Followers
Vivek retweeted
good luck paying $20k only to find out you can generate 20 tok/s. even running 24/7, that's just 50m tokens/month. for glm, at $4.40/m, this is $228 in value. any $200 sub gives you significantly more. and this math means the break-even is 7.3 years not 6 moths. by that time the hardware would die if it's running 24/7.
Jeffrey Scholz@Jeyffre
1 - So GLM 5.2 is 700b parameters (ish) 2 - 4x DGX Sparks can supposedly handle up to 700b parameters (give or take) 3 - GLM 5.2 is supposedly in striking distance of the performance of GPT 5.5 and Opus 4.8. In my brief tests, it's really not shabby at all. 4 - So for $20k, you can get near the frontier on your table. 5 - Extrapolate the trend, and you could have mythos/5.5 pro - class models in your dining room for the cost of a cheap car less than five years from now. Even without extrapolation, we're already the near frontier running locally. 6 - Paying real api costs, I could easily blow through $3,000 per month coding and running agents. The machine pays for itself in 6-7 months conservatively. 7 - In 3-5 years, most power users of AI will self-host. 8 - Am I missing something?
English
Vivek retweeted
Vivek retweeted
🚨 Crude Oil is Almost Back to Pre-War Levels
Petrol Was Hiked 4 Times by ₹7.5 in 10 Days
Why Aren't They Quick to Decrease the Prices?
Benefits Should Be Passed To Taxpayers Now
English
Vivek retweeted
Vivek retweeted
Vivek retweeted
BREAKING: Iran has declared that ships cannot cross the Strait of Hormuz without its permission, setting the stage for future tolling arrangements by saying it could introduce “insurance fees," per Bloomberg.
Details include:
1. All vessels that transit the Strait of Hormuz will have to secure a mandatory insurance policy
2. This policy is currently free for the next 60 days but could involve charges in the future
3. Iran also said ships must follow its prescribed route and that alternatives are prohibited
The MoU signed with the US only says that transit through the Strait of Hormuz would be free for the duration of its 60 day term.
It appears Iran is preparing for long-term control of Hormuz.
English
Vivek retweeted
Credit where it’s due.
A group of pro-India patriots on X managed to spark a massive conversation around issues China would rather keep hidden. The campaign’s impact was so significant that even Chinese state-linked English media outlet Global Times is now issuing clarifications 🤣
Whoever was behind it absolutely nailed the execution🫡👌🏾
Global Times@globaltimesnews
Recently, some Indian social media users claimed that they had made a "discovery" that China, like India, supposedly has "a caste system." They shared memes about ancient China's "four occupations" (Shi-Nong-Gong-Shang: scholars, farmers, artisans, merchants) by framing them as equivalent to India's hereditary jati system. Chinese commentators believe the claim reflects the ignorance of some Indian netizens about Chinese historical culture. globaltimes.cn/page/202606/13…
English
Vivek retweeted
#BREAKING: New diplomatic standoff between United States and Italy after US President Trump claimed Meloni begged him for a photograph at the G7 in France.
Meloni has accused her one-time close ally Donald Trump of fabricating the story about her.
English
Vivek retweeted
China got brutally exposed. Their state media, Global Times, is failing to defend them. The only reason Congress supporters are still defending China even without getting paid is because Nehru is the only PM in Indian history who has lost a war. So eventually, Congress supporters have to claim that Chinese avengers are 1,000 years ahead from the rest of the world 😂 😂 😂
English
Vivek retweeted
Vivek retweeted
🔥 NEW: Claude Code has launched Artifacts, enabling users to share interactive apps and dashboards via private links.
English
Vivek retweeted
Vivek retweeted
Vivek retweeted
Folks who are running GLM 5.2, how are you doing it?
What harness/provider are you using?
Getting FOMO about an open weights model for the first time
English
Vivek retweeted
பீகாரின் பெகுசராய் மாவட்டத்தில் 30 வயது பெண் ஒருவர் ஐந்து பேர் கொண்ட கும்பலால் கடத்தப்பட்டு, கூட்டுப் பாலியல் வன்கொடுமைக்கு ஆளாக்கப்பட்டதோடு மட்டுமின்றி, மிகக் கொடூரமான முறையில் சித்திரவதையும் செய்யப்பட்டுள்ளார்.
பாதிக்கப்பட்ட அந்தப் பெண்ணின் உடலில் இருந்து துப்பாக்கித் தோட்டாக்கள், கற்கள் மற்றும் மரத்துண்டுகளை மருத்துவர்கள் கண்டெடுத்துள்ள அதிர்ச்சித் தகவல் தற்போது வெளியாகியுள்ளது.
தமிழ்
Vivek retweeted
We need more benchmarks like DeepSWE. DeepSWE properly puts GLM 5.2 close to Opus 4.8 and GPT 5.5.
English
Vivek retweeted
GLM 5.2 is a massive jump from GLM 5.1.
But there is a tradeoff nobody is talking about.
You will prompt GLM 5.2 three or four times to accomplish what GPT 5.5 or Opus 4.8 nails in a single prompt.
Open source models are missing a level of intelligence that benchmarks do not measure.
Call it taste, instinct, attention to detail.
Whatever it is, the closed labs still have more of it.
But here is the shift.
We are entering an era where cheaper open source models can handle most real work.
Fable 5 was overkill for 80% of vibe coding tasks anyway.
The US government gave Z.ai this opening by banning Fable 5.
GLM 5.2 walked right through it.
Unban Fable 5.
English
JUST IN: OpenAI and Anthropic push customers toward token based billing as operation costs rise.
English
Vivek retweeted
Great Move by Tamil Nadu Government! 👍
For the first time, the personal contact numbers (along with official emails) of all Ministers are now publicly shared on the official website. This is a fantastic step towards transparency and making governance more accessible to the people.
Full Council of Ministers list with contacts:
👉 tn.gov.in/minister_list.…
Transparency like this should be the norm!
English
Vivek retweeted
GLM-5.2 delivers a substantial leap in app development capabilities, which also represent demanding long-horizon tasks.
Results:
- GLM-5.1: 21/70
- GLM-5.2: 48/70
- Claude Fable 5: 56/70
That's more than a twofold improvement from GLM-5.1 to GLM-5.2.
These come from an internal benchmark of 35 challenging mobile development tasks, each run twice for a total of 70 trials. We measured task completion, defined as core features working without major issues.
English