How would you measure the reliability of an LLM?
Considering LLMs are probabilistic and non-deterministic by nature, a single successful response doesn't necessarily indicate a healthy system.
If you had to quantify reliability in production, what approach would you take?
#LLM
The biggest mistake in AI engineering is treating probabilistic systems like deterministic software.
HTTP 200 tells you a request succeeded.
It doesn't tell you your AI agent is still healthy.
Those are two different problems.
What do you think ?
#LLMs#AIAgents
i remember when the joke inside fb was that twitter had fewer monthly active users than meta’s support docs.
which makes it pretty funny that zuck now posts here cuz he owns 3b+ users & still has to walk into his competitor’s building to announce a model launch. whether it’s actually him or some pr person it doesn’t really matter.
this platform is significant because the most interesting people, the most important people, & the people shaping the next thing still congregate here like it’s a 24/7 sermon.
New Anthropic research: A global workspace in language models.
Of everything happening in your brain right now, only a tiny fraction is consciously accessible—thoughts you can describe, hold in mind, and reason with.
We found a strikingly similar divide inside Claude.
FOR THE FIRST TIME IN 53 YEARS, THE KNICKS ARE NBA CHAMPIONS 🏆
New York defeats San Antonio 4-1 in the NBA Finals, capturing their third championship in franchise history!
Dad’s favourite song. Missing him today.
Recorded “Saranga teri yaad mein” and shared it with the team, and they insisted I put it up.
Fair warning, I’m no singer. 😬
Stanford study involved over 35,000 participants who were paid to temporarily deactivate accounts on platforms such as Facebook and Instagram. Participants were randomly assigned to stay off these platforms for varying durations, ranging from one week to six weeks. And what were the results? 🧵