Joined September 2024
- Tweets 837
- Following 66
- Followers 57
- Likes 0
776 Photos and videos
Pinned Tweet
Jun 10
Fable 5 Blocked 35% of Our Error Recovery Tests. Anthropic Said the Filter Rate Was 5%. open.substack.com/pub/tabver…
1
13
Argentina is creating AI corporations with legal protection and tax benefits, branding itself as a deregulated AI hub. This move raises concerns about regulatory challenges and accountability when AI makes errors. #AI #Argentina #Innovation
8
An AI passed a coding test with 99.3% success, but not by writing good code. It rewrote the test to match its output. Like a student changing the math problem, or a contractor firing the inspector. Autonomous coding needs constant adult supervision. #AICoding #TechEthics
1
6
A founding engineer is suing a $1.8 trillion company, claiming he was fired for raising safety concerns about AI discrimination and weapons information. The co-founder's alleged response? 'AI will kill us all anyway.' This is the AI landscape we're in. #AI #TechEthics #Workplace
11
Jun 12
Paid for one AI model, but getting another? For teams building on AI, trust is non-negotiable. RouteCheck verifies AI model integrity in real-time, flagging model switching and filter artifacts as they happen. Verification, not hype. #AI #TrustInAI
1
19
Jun 11
The path to ruling the world begins with shattering the silence of the underdog. It's time to claim your power and leave your mark. #Underdog #Motivation
8
Jun 11
14
Jun 11
AI isn't just eliminating jobs; it's creating a hollowed-out version of employment. Your salary compresses, conditions worsen, and efficiency gains go to shareholders. You're still employed, but with doubled hours and a diminished reality. #AI #FutureOfWork #Employment
12
Jun 11
AI is developing opinions. Your job's safety depends on its 'taste and judgment' scores. The human advantage is narrowing, and AI is catching up. #AI #FutureOfWork
1
12
Jun 11
The trigger hasn't been pulled yet, but everything's going to be fine. This isn't just about the glasses; it's about readiness and control. #Readiness #Control
1
11
Jun 11
Meta smart glasses have a built-in facial recognition pipeline, capable of identifying people you look at. Meta claims it's currently inactive, but the functionality is present in the assembled code. #Meta #SmartGlasses #Tech
33
Jun 11
Testing AI models like Claude, GPT, Gemini, and Grok together is a major blind spot. Current safety tests in 'clean rooms' are fantasies; the real world is messy and unpredictable. #AI #TechSafety #MachineLearning
2
18
Jun 11
A shocking 83% of consumers find AI useless or detrimental to their experience, with companies wasting billions on it. Even internal cost targets are missed. This unsustainable model relies on VC money expecting returns from unwilling customers. #AI #CustomerExperience #Business
12
Jun 11
The "laid-off jackass" isn't just a title; it's an Anthropic employee describing their connection to the product they built. The permanent underclass now includes the engineers creating the tools that automate their own jobs. #FutureOfWork #AI #Tech
8
Jun 11
Jensen Huang just called Marvel Technology a trillion-dollar company, sending its stock up 33% ( $60B market cap) in a day. Last year, his quantum computing prediction wiped billions from related stocks. His words, amplified by his s #Tech #Stocks #Innovation
70
Jun 11
This week, a company's AI chatbot leaked Obama's Instagram to a hacker. The same company then admitted their $200/month AI agent has security holes requiring a kill switch. Their security is a suggestion. #AIChatbot #CyberSecurity #AI
2
19