This is consistently the most effective newsletter for catching up on important news in tech and science--and witty to boot (LOL to "making the county the second Grimes to bear Elon’s children.").
I am confused. Why is it necessary for more data centers when agents will soon be able to run on your laptop and phone without being connected to the internet or data centers? Desktop Agent Computer
Powered by the NVIDIA GB10 Grace Blackwell Superchip, NVIDIA DGX Spark™ is a complete platform for local autonomous agents. Large local memory, powerful AI compute and the NVIDIA AI software stack enable agents and large models to run locally, reducing the need for cloud-based token generation resources. With a compact, power efficient design, DGX Spark is built to run always-on agent workloads - right from the desktop.
Exponential need / demand … biology for example may need quantum computers to solve longevity- the complexity is far beyond what we comprehend… imagine having a tracker / scanner on or in your body learning and tracking everything in your body - reading it / analyzing it forecasting and all - and there are exponentially more fields and opportunities we don’t even understand yet
Yes, and data centers are not needed for any of these. I am about to release my 1st novel that mentions an internal AI device that tracks everything you do, can even hear your internal thoughts, and can see what you see and smell. I am having a pre-sale for the eBook now until September 1st, and then the physical version will be released. I won't put a link here, but if you go to Amazon and search my name or the title of the book, you can find it easily. I also have a short reading on my Substack. It is called "The Love Boat 2120". The current chips for the current data centers are obsolete, by the way, and the next generation is 1000X more efficient and powerful, using superconducting materials that AI has helped to find. No copper needed means no excess heat and the chips can be stacked without burning up. A refrigerator-sized stack as powerful as a whole data center!
Amazing and “brave” - so many potential futures ahead - one thing I think is that while the speed of AI growth will be unreal - physical AI will be slower — building a billion robots human or not is gonna take some time - and the new opportunities will be enormous before the “retirement” for most humans … searching for the book now :)
Thanks Erik. I need all the help I can get. My novel is a bit different and unlike most science fiction today, which is dystopian, it is a positive view of the future.
I've been enjoying the conversation between you two. It seems to me that the AI will be so varied and compute hungry that we will well use the local compute and for truly civilizational level tasks mega-data centers will be employed. Compute will get cheaper and more compact AND we will need more if it. A Jevon's paradox thing.
Good luck with the book + and -. I'll check it out as well.
Thanks Jesse, I agree and believe the data centers will not be able to keep up. The compute solution will be qubits expanding to 100,000 in quantum computers. I think the first 100 qubits is now created! When that will be achieved is the question. The most complex problems will need this level of compute, such as worldwide control of weather and climate and the ending of all wars.
The true value of your daily podcast as I see it, extends far above than its viewership, likes, comments, or audience size. Its greatest significance lies in its future historical and intellectual value, a body of thought that may endure as a lasting record and emblem of human intellect, perhaps even beyond the age of singularity
Hey AWG, thanks for your newsletters! I am still quite young, and so far, I don't have much of a technical background yet. There's still a lot I don't know and many concepts I've never encountered before. Also, I am not a native English speaker. However, your newsletter challenges me in a way that makes me want to learn more, broaden my horizons, and understand English used in contexts and styles I haven't seen before. When I read your newsletter, I have to think. When reading other newsletters, I just read. Thanks!!
My first job out of college was as an engineer for 8 bit microcontrollers that went into auto safety applications. We would hard code our customer's software into silicon, creating a more robust system for the variety of environments cars encounter. Taalas's central idea is a blast from the past.
Alex, 🙏 thank you for your efforts. I’d love to hear your sense of the math of investing in the F500 (98% not Tesla, 100% not spacex) — when the trend has been for the biggest winners to continue to win. There will be amazing companies that come up. There will be some winners and many losers in the S&P. But it would seem to be that what Elon is building warrants some weighting, no?! How do you think about the math for this if you cared to maximize your position?
Stay up tonight in hopes of another compelling Welcome To? Of course! As a square, the completed Terafab would reach 1.89 miles per side, 46X my 50 acre Midwest hobby farm!
You are outstanding, Alex. Thank you for everything you do, your insights, analysis and just overall brilliance. Onwards!
Thank you AWG👏🏽
This is consistently the most effective newsletter for catching up on important news in tech and science--and witty to boot (LOL to "making the county the second Grimes to bear Elon’s children.").
Working on the solve everything singularity alongside others in earnest. Thanks for your work and inspiration.
I am confused. Why is it necessary for more data centers when agents will soon be able to run on your laptop and phone without being connected to the internet or data centers? Desktop Agent Computer
Powered by the NVIDIA GB10 Grace Blackwell Superchip, NVIDIA DGX Spark™ is a complete platform for local autonomous agents. Large local memory, powerful AI compute and the NVIDIA AI software stack enable agents and large models to run locally, reducing the need for cloud-based token generation resources. With a compact, power efficient design, DGX Spark is built to run always-on agent workloads - right from the desktop.
Exponential need / demand … biology for example may need quantum computers to solve longevity- the complexity is far beyond what we comprehend… imagine having a tracker / scanner on or in your body learning and tracking everything in your body - reading it / analyzing it forecasting and all - and there are exponentially more fields and opportunities we don’t even understand yet
Yes, and data centers are not needed for any of these. I am about to release my 1st novel that mentions an internal AI device that tracks everything you do, can even hear your internal thoughts, and can see what you see and smell. I am having a pre-sale for the eBook now until September 1st, and then the physical version will be released. I won't put a link here, but if you go to Amazon and search my name or the title of the book, you can find it easily. I also have a short reading on my Substack. It is called "The Love Boat 2120". The current chips for the current data centers are obsolete, by the way, and the next generation is 1000X more efficient and powerful, using superconducting materials that AI has helped to find. No copper needed means no excess heat and the chips can be stacked without burning up. A refrigerator-sized stack as powerful as a whole data center!
Amazing and “brave” - so many potential futures ahead - one thing I think is that while the speed of AI growth will be unreal - physical AI will be slower — building a billion robots human or not is gonna take some time - and the new opportunities will be enormous before the “retirement” for most humans … searching for the book now :)
Thanks Erik. I need all the help I can get. My novel is a bit different and unlike most science fiction today, which is dystopian, it is a positive view of the future.
I've been enjoying the conversation between you two. It seems to me that the AI will be so varied and compute hungry that we will well use the local compute and for truly civilizational level tasks mega-data centers will be employed. Compute will get cheaper and more compact AND we will need more if it. A Jevon's paradox thing.
Good luck with the book + and -. I'll check it out as well.
Thanks Jesse, I agree and believe the data centers will not be able to keep up. The compute solution will be qubits expanding to 100,000 in quantum computers. I think the first 100 qubits is now created! When that will be achieved is the question. The most complex problems will need this level of compute, such as worldwide control of weather and climate and the ending of all wars.
Dear Dr. Alex,
The true value of your daily podcast as I see it, extends far above than its viewership, likes, comments, or audience size. Its greatest significance lies in its future historical and intellectual value, a body of thought that may endure as a lasting record and emblem of human intellect, perhaps even beyond the age of singularity
Thank you for your outstanding work.
Hey AWG, thanks for your newsletters! I am still quite young, and so far, I don't have much of a technical background yet. There's still a lot I don't know and many concepts I've never encountered before. Also, I am not a native English speaker. However, your newsletter challenges me in a way that makes me want to learn more, broaden my horizons, and understand English used in contexts and styles I haven't seen before. When I read your newsletter, I have to think. When reading other newsletters, I just read. Thanks!!
I love the non-explanation you've given to the Pangram overseers. Did anyone else notice it? "We are borg".
I had to stop and let those Terafab numbers sink in for a moment. Just wow.
Funny about Grimes and Growth and Beauty
My first job out of college was as an engineer for 8 bit microcontrollers that went into auto safety applications. We would hard code our customer's software into silicon, creating a more robust system for the variety of environments cars encounter. Taalas's central idea is a blast from the past.
Alex, 🙏 thank you for your efforts. I’d love to hear your sense of the math of investing in the F500 (98% not Tesla, 100% not spacex) — when the trend has been for the biggest winners to continue to win. There will be amazing companies that come up. There will be some winners and many losers in the S&P. But it would seem to be that what Elon is building warrants some weighting, no?! How do you think about the math for this if you cared to maximize your position?
Stay up tonight in hopes of another compelling Welcome To? Of course! As a square, the completed Terafab would reach 1.89 miles per side, 46X my 50 acre Midwest hobby farm!