Aadhar Chandiwala’s Post

This is something humans should learn from Machine Learning Algorithms 👇 Don't get stuck on Local Maxima. Always experiment and explore in search of the global maxima!

Banks and insurance companies hide a lot of information from their users through jargon-filled language. Isn't someone building an AI startup yet to uncover all these hidden terms and conditions?

3-4 year old kids are being sent to coaching sessions after their preschool classes are over so that they can CRACK interviews for good schools!!! 🤯 Toxic

Innovation at this pace, when you have a Listed Company. What a Leader! 💯

Building support for A/B tests in the product might be a one-off investment, but it gives much better velocity on experimentations, at least twice as fast, as making sequential changes 🚀

Almost all the awesome developers I know believe that coding is as much an art as it is a science! 🖌️

Back in the 1950s, High level programming languages were designed to address human weaknesses in understanding and writing assembly code effectively. And, LLM’s are just another progression towards making it even simpler for humans to write logic. Who knows, someone might be building a robust natural language programming language that compiles directly into a low-level language, making it fast as well as easily debuggable💡 Let me know if something like this getting built!

I generated this animation on "How Context Switching Happens in a CPU" by prompting Claude Opus! ⚡️ I used Claude to write a script in Python's Manim package, while this lacks some nuances of the concept, it did a better job than any of the LLMs out there, including ChatGPT. What do you think? 💡

This is 💯 percent on point! AI companies have found a great way of getting viral using these "walkthrough" videos, which are hardly close to reality. For instance, Google's release of Bard's walkthrough video vs. the reality, and we all know it sucks! Sora looks promosing as well on demo videos, but who knows what’s the reality

🚨All versions of Java 8 to early versions of Java 22 unexpectedly keep crashing if you update to macOS 14.4, and there is NO workaround yet! This is because during the JIT (Just-In-Time) compilation, some processes may access protected memory regions. Before macOS 14.4, for such memory accesses it used to send a SIGBUS or SIGSEGV event, and the process chose to handle it gracefully. However, after the update, such memory access events are sending a SIGKILL event, leading to unconditional termination. Link to the official Oracle blog post and Java Bug Report in the comments 👇🏻