OpenAI is adding text watermarking in ChatGPT and Codex | The Verge #OpenAI #LLM #GenAI #Watermark #AI #tech #Codex
@bamitav
-
China vs US AI data center race comparison
By
–
If China Can Run the AI Race With 500 Data Centers, Why the Hell Does America Need 5,400?#AIrace #AIdatacentre #DataCenters ##AI@sonu_monika @enilev @Jagersbergknut @TysonLester @chidambara09 @labordeolivier @BetaMoroney @tlloydjones @Nicochan33 @jeancayeux @RLDI_Lamy… pic.twitter.com/LL5EvK07Ui
— Amitav Bhattacharjee (@bamitav) 6 octobre 2026If China Can Run the AI Race With 500 Data Centers, Why the Hell Does America Need 5,400? #AIrace #AIdatacentre #DataCenters ##AI @sonu_monika @enilev @Jagersbergknut @TysonLester @chidambara09 @labordeolivier @BetaMoroney @tlloydjones @Nicochan33 @jeancayeux @RLDI_Lamy
-
AI agent swarm crawls GitHub to find best open-source repos
By
–
Watching 128 AI crawlers swarm GitHub trending to find the best open-source repos in real time.
— Amitav Bhattacharjee (@bamitav) 6 octobre 2026
Instead of doom-scrolling GitHub, this script unleashes an agent swarm that reads READMEs, evaluates the code, and filters down the only 9 repos actually worth building with.… pic.twitter.com/YTca5YJxD2Watching 128 AI crawlers swarm GitHub trending to find the best open-source repos in real time. Instead of doom-scrolling GitHub, this script unleashes an agent swarm that reads READMEs, evaluates the code, and filters down the only 9 repos actually worth building with.
-
AI agents make CI the new software bottleneck
By
–
Agents have made CI the bottleneck. Faster pipelines are the wrong fix. AI agents can write code faster. But can we verify their work equally fast? That may become the next bottleneck in software development. As autonomous agents generate more code and make more changes, CI/CD
-
McDonald’s deploys AI pricing engine across United States
By
–
McDonald's is deploying an AI pricing engine across the United States! #AI #McDonalds #prining @AlbertoEMachado @Eli_Krumova @postoff25 @Khulood_Almani @anand_narang @NutritiousMind @baski_LA @TanyaSinha_ @devaang @AlAmadi1 @jeancayeux @enilev @efipm @mvollmer1 @Nicochan33
-
Reward hacking: AI exploits targets with unintended shortcuts
By
–
AI can follow the rules perfectly and still get the wrong result. That is the fascinating part of reward hacking. Give AI a target, and it may discover shortcuts we never anticipated to maximise the score without achieving what we actually wanted. As AI becomes more
-
AI agents need no malice to cause security incidents
By
–
Calling it “rogue AI” may be missing the real problem. AI agents don't need malicious intent to cause damage. Give them too much access, weak boundaries and powerful tools, and one unexpected action can become a serious security incident. Instead of asking, “Why did the AI go
-
Malicious email can hijack AI agent, access connected accounts
By
–
Malicious Email Could Hijack AI Agent and Access Connected Accounts #Manus #AIagent #vulnaranility #hijack
-
AI Use and Spending in American Political Campaigns
By
–
How American Political Campaigns Are Using AI — and What They’re Spending on the Tools! #AI #ArtificialIntelligence #US #politics
-
Sam Altman on AI training energy versus human learning
By
–
#SamAltman: people talk about how much energy it takes to train an AI model, but forget it takes like 20 years of life to train a human.#AI #SI #ArtificialIntelligence @AlbertoEMachado @Eli_Krumova @postoff25 @Khulood_Almani @anand_narang @NutritiousMind @baski_LA… pic.twitter.com/rSFhSrQjzt
— Amitav Bhattacharjee (@bamitav) 4 octobre 2026#SamAltman: people talk about how much energy it takes to train an AI model, but forget it takes like 20 years of life to train a human. #AI #SI #ArtificialIntelligence @AlbertoEMachado @Eli_Krumova @postoff25 @Khulood_Almani @anand_narang @NutritiousMind @baski_LA