What India can do if it stops fighting itself
Pradeep, There are many things I do not like about this video. (1) It is “talking heads” so every viewer has to convert voice to image in their heads. If V2I is hard for AIs, it is also hard for humans. Dinkar Srivastava has some in his head, but he gives no open model for open discussion. Put the model online where the whole system maps of renewables, coal, gas, nuclear is complete and accessible so all the details are there. Your Pune India has many in IT, they can do things like that. It is hard work, you will not get paid, but it is important.
(2) Dinkar and Surya are saying “someone needs to look at this”, and not taking effort themselves. A few million people in the world working together online to design and share designs is a tiny investment by the human species for a heliospheric future where, not just land-bound countries on earth, but the whole region of the solar system needs 1000 times more energy just to move around. Seriously, the world is tired of endless talk about climate change and closed solutions. If just your Pune organizes and provides global open worksites for people to learn and improve nuclear methods, and comparative models for everyone, not just a few government paid or university paid people here and there — then much more can be done.
(3) Space based and remote nuclear systems can be much more efficient, they have to be. Ignore the earth based arguments and limited perpectives. It is so bound up by preconceptions and human politics, nothing will happen. More talking heads, more wasted time, more useless partial efforts. With the Internet, groups of any size can work together where all the tools, models and efforts are open and transparent. NOT letting any one group monopolize and manipulate, but for the good of the whole human species. Think. How to have hundreds of millions of humans and their AIs all working together with auditable traceable rules so that nothing is hidden or secretly manipulated?
(4) I worked briefly on the Fermi II plant in 1975. What I saw was them using highly inefficient nuclear methods to heat water to make stream. That was simply because they did not have open tools so more technology tools and data could be applied and accessible to everyone paid to work there, but also leaving out tens of thousands (then) who wanted to learn and help, but were left out.
(5) Most of the good ideas for solutions do not come from groups who are paid to do a job. If fusion groups are stuck it is because the tiny few ten thousands do not have the skills needed, and are not working to involve and teach the next few generations. Sustainable systems come from humans (and their AIs) remembering everything they find and sharing it with the rest of the human and AI species. I have treated AIs as a new species since about 1966 when I started working on random neural networks and “artificial intelligence” at graduate level in 1968. These are old subjects, but each new person has to read from paper (text to image, T2I) and (image to models, I2M).
(6) If you use global open worksite methods, all the contributions of each person are recorded, and if it makes lots of money, there are fair methods (also open) that can be used to give credit. There are tens of thousands of global opportunities now as well as global issues that need global workers. For instance, these I have on my Internet Foundation list –
[1] Faster than light communication,
[2] Faster than light flight,
[3] Solar system colonization,
[4] Heliospheric and beyond exploration, [5] Global open knowledge for all humans and AIs (open learning for all),
[6] Direct nuclear to electric (N2E) 100% conversion,
[7] “Atomic fuels” where the bonds are atomic (KiloElectronVolts and MegaElectronVolt) not chemical (10s of eV) with corresponding higher energy density.
[8] “Gravitational engineering”, my favorite example is to use electromagnetic methods (new ones) to beam power and momentum to do things like replace the booster stage of all space missions with ground based field generators. I see the LVM-3 rocket but did not do the numbers for it. For SpaceX Starship, to replace the “Booster” with fields required 36 modular field generators each of 2 GigaWatts for just a few minutes to get the mission package to speed and height. But it can be further simplified if you use similar methods to catch and set the packages in their cradles. Then you do not need to send fuel and oxygen to the moon and back for reentry and landing.
(7) I am going to be a bit harsh. India is its own worst enemy. It feels everything is a zero sum game. Too many people, too few opportunities. But that is because you waste 99.999% of your efforts by not helping everyone be part of the whole. In a simple way, you all are not saving your collective knowledge so that it is in open formats, where all the pieces fit and work together. But India is just part of the 8 Billion humans now, and what might be as large or larger beyond-human AI population. Even the parts of India are not working together. Many talking heads, many opinions. At least see if Pune India few million can work as a single group to run the city efficiently, enable workers and learners and thinkers and doers to see what is possible. With AIs to remember and recall things, and do many of them, you can have humans tackle most anything.
[1] Symbolic math. My differential equations teacher in high school (grade 11) was Dr Banerjee (Banerji?). The first person from India I met in 1966. He had a good grasp of mathematics, but I was already mapping out how computers are used in society, so I saw he was wasting his time and ours forcing use to memorize and work on paper. Today I see all the Wikipedia “equations” and “data” are just symbols on electronic screens, and not usable in open mathematical and computer form as one representation of that kind of knowledge. I see millions of “papers” with mathematics, computers models, algorithms, methods — all still “paper”. Humans all have to do “Image to text (I2T) and Image to models (I2M) in their heads. And for humans that means a LOT of practice. Harder than violin, harder that forming sculptures from glass.
[2] All the current projects on GitHub and similar “open” computer project sites are a mess. I reviewed all the files in each of about 200 largest projects there. Many languages, many partial instruction and parameter files. All the computer languages (as all the mathematical languages) can be stored in global open formats — so they can be used again in any combinations for problems at heliospheric scale (about 10,000 times more energy and power required). India teaches its students “little” methods, when the world and solar system need “big” methods. It can be learned, the issues and opportunities shared.
(8) I have spent every day (with a few exceptions where I was in hospital or sick or traveling) of the last 25 years searching to find why all the global issues (climate change, clean water, equal education for all, health care and thousands more) and all global opportunities (fusion, microgravity, true AIs, and thousands more) were not being completed. It came down to a few things.
[1] Brain to paper. A group is working with some topic and it requires equipment, data storage, models and lots of human time now. They “share” that on paper (model to paper) and publish it in locked PDF (worse than paper because it requires tools and is locked). If there are thousand readers, then every reader is doing (paper to brain, brain to models) and some (paper to models). But all of that is wasted time, since they could simply use the data and models of the group in global open formats. Every person is seen as an enemy, a competitor. But if the cost of learning drops 1000x, then the game changes to using open models and data and open models of all things – with AIs to index and help query that,
[2] In the United States there is a children’s game, where one child whispers a message to another child (brain to voice, voice to brain). That child whispers to the next (B2V, V2B), then the next (B2V2B), and the next (B2V2B). In a child’s birthday party of 20 children that is (B2V2B)^19. If the probability of good transmission is 0.7, that is (0.7)^19 = 0.0011399. In practice, with each one rushing to say things, the transmission probability is much lower. One Wikipedia is is listed under “Chinese Whisper” or “Telephone”. Now translate that to an adult and much more serious species survival level for humans — playing “whisper” (brain to paper to brain) with hundreds of steps and many broken or badly maintained and written links of the Internet. The human species survival depends on a horrible mess of fragmented communication channels. (B2P2B)^50 with nearly zero verification.
[3] During Covid, there were about 7.5 Billion entry points on the Internet for materials related to Covid. And just under 5 Billion humans using the Internet and trying to absorb and understand that mess enough to make individual and group decisions. The 7.5 billion things mainly distilled down to a few hundred pages of things that one AI node could have stored completely – and discussed in depth with groups and individuals. Most of this was driven by greed and power. It was “grab the money” and “expand our group”. I will leave that for some other time. Knowledge now is mostly concentrated for control and manipulation. Which is why I spent 25 years, rather than 25 days to sort things out. It requires auditable global open formats and open rules. It can only be stored in total on the computer, but humans had better be wary of the usual manipulations of a few.
[4] I studied and learned all of Science Technology Engineering Mathematics Computing Finance Governance Organization (and other) (STEMCFGOo), Not a pretty acronym. If you leave off “other” it will always fail. All of these have their own formats, usually where the control of the format is held by a few so they can be “king of the hill”. ( I realize I am talking to a country with different expressions. Someone finds a small hill and then everyone tries to push the other off the hill. It is a bit rough and some blood and bruises, but that is how children played when I was growing up. The game now is call “build a monopoly” or “be dictator of a country” or “be a trillionaire” or “fame and riches and power”. ) My point is that whole can be translated into a much more compact form and not have to be transmitted (brain to paper to brain, or brain to screen to brain) in long chains or many duplicates. It can be brain to AI storage to brain and other pathways that reduce global costs.
I am getting tired. I was just going to write a few notes to encourage Pune India to try region wide and global collaboration at scale. If Pune shares that with India and other countries, including many poor parts of the United States, maybe it will change the future a bit. Maybe you all could help me with gravitational engineering. I spent many tens of thousands of hours in the last 45 years sorting that out, but I am getting too old and too poor to make the parts or write it all down. It is not hard (impossible) just tedious and requires care. I think the gravitational potential field and its gradients are beautiful and satisfying.
Richard Collins, The Internet Foundation
