deeznuttynutz

Gee Golly!! We are accelerating towards the Singularity at BREAK NECK pace! I hope Garry can accelerate and increase his LOC to 200K per day so we can get more blog and hopefully AGI soon!

I guess TPUs and JAX give it a third chance, and maybe MLX a fourth, lmao.

This is an interesting model. I have an embodied AI companion app on the App Store that uses Gemini 2.0 Flash. However, I've been searching for a model that is comparable to 2.0 Flash in cost with more creative and less repetitive outputs. The frontier labs have only gone up in cost and latency, and are ignoring use cases like this. This seems promising, I plan to try it in production.

That's not even close to remotely true. If that is the case, why is the UAE throwing their money at US startups and not their own? They have more than enough "capital".

Agreed, it's not that bad imo and the profiling tools are amazing.

It's a pity?

Please explain in detail how alternatives would have worked better for GCC and CUDA. Also, if you could share some historical context about how those alternatives could realistically have been implemented at the time, that would be helpful too.

I love to hear all the "would've" and "should've" scenarios.

Well, there is an app called Tolan (alien AI friend) that has been very successful and the devs have said that 80% of the users are young women.

I myself have released an app in this realm a few days ago, it's very much a work in progress, but my goal was to let the AI feel more like a computer and less like a companion/boyfriend. I think the relationships these companies are pushing will be harmful in the long term.

My app if you are curious: https://apps.apple.com/us/app/lonely-bots-3d-ai-friends/id67...

This is exactly while I'm building my app now with the expectation that these assets will be exponentially better in the short term.

Of all the recent books I've read about AI, this was by far the worst. The Singularity is Nearer, Life 3.0, and A Brief History of Intelligence were much much much better imho

Crazy...

Quantized Inertia is a new theory that works and is interesting. Although, not rigorous and is speculative at best

This 1000%

I completed the book Programming Principles and Practice using C++ (which I HIGHLY recommend to any beginner interested in software engineering) about year ago with GPT4 as a companion. I read the book throughly and did all the exercises, only asking questions to GPT4 when I was stuck. This took me about 900-1000 hours total. Although I achieved my goal of learning C++ to a basic novice level, I acquired another skill unintentionally: the ability to break down tasks effectively to LLMs and prompt in a fashion that is extremely modular. I've been able to create complex apps and programs in a variety of programming languages even though I really only know C++. It has been an eye-opening experience. Of course it isn't perfect, but it is mind blowing and quite disturbing.

We have lived long enough to see Physics become Modern Math and create no new provable Physics

What's the good word!!

I'm similar to you in many regards. I had no desire to learn math early in school due to the education system, how it is taught and lack of meaning for math portrayed early on. I coasted through high school like a zombie without meaning until 11th grade when I took Physics. Suddenly everything clicked, I magically became good at math despite not performing well at it before in my math classes. I ended up studying engineering and working in the EV industry. Now, I am studying pure math for the sake of my own curiosity and I'm passionate about developing AI tools to help people learn and see the why behind math as early as possible. I think I would have accomplished much more if I was exposed to Physics in elementary or middle school or at least a "History of Math" philosophy based class.

I've read it and think it gives great historical perspective to the field. It is important that one understands the successful/unsuccessful methods and subfields developed since Turing's paper. IMO it is hard to appreciate the modern advances without the big picture. Some of the classic symbolic techniques combined with modern ML architectures may come full circle towards solving AGI in the way that NN came back to become the foundation of machine learning.

FF7 style?

All that is missing is the "Xbox E", hopefully the next one will be called that. Let's pray we get a special Tesla edition on launch day.

Over the past year, I've completed most of the second edition. It truly is an amazing book that has helped me overcome many mental barriers I've faced with programming for years. The main reason for the reduced size of this edition is the removal of 'Part IV: Broadening the View.' This section, which covers additional topics such as text manipulation, numerical computing, and embedded systems, is now available online. The chapters have not been updated from the second edition, as the topics are still relevant and utilize C++11/14.