« ATOM Award of the Month, November 2020 | Main | ATOM Award of the Month, January 2021 »

Comments

Drew

Biology is the one partial exception along the lines of time for atom assembly so far - there is that old project management joke that you can't get a baby by hiring 9 women for 1 month.
We are definitely seeing technology's impact on understanding biology so maybe we will crack that last area. A lot of the descriptions of "magic" nanotechnology is assuming a world in which that is true and a body can be created/modified quickly.
Precision fermentation also does it to some extent and might be one of the examples of the ratcheting of what we consider futuristic technology vs conventional technology - just like we don't think about our phone giving us directions as artificial intelligence we don't think of microbes producing human hormones as faster biology.

Kartik Gada

Drew,

Computational power has accelerated the speed of genomic and proteomic research, even if it seems to have stalled for several years. This is due to the limitations of Von Neumann architecture vs. more advanced (Quantum) approaches to computing.

The innovation from Moderna to create the Coronvavirus vaccine is also a radical new innovation, that brought vaccine discovery closer to software-like principles.

Geoman

Thanks for the unexpected kudos.

Biology is the exception, but it really isn't. DNA is a repository for an immense amount of information. It takes nine months to make a baby because that is the download time for all the biological information necessary for new life. Think of it as copying a really large hard drive from one computer to another. And proofing and correcting all that data.

Type the words "Biology" and "cracking the code" into google. You'll see an endless number of articles there that use those terms. That is not just hyperbole - it is exactly what they think they are doing, trying to crack an information code. Hacking an unfamiliar computer system. It has proven to be incredibly difficult. When Watson and Crick solved the structure of DNA, we thought, that's it. All we did then was find the hard drive on the computer.

Imagine an alien supercomputer computer fell from the sky. At first we can just see it function. We can intuit some of how it works by measuring inputs versus outputs.

Slowly we start to take it part, examining components. We can figure out what each does, but not how it works as a whole.

Finally we understand all the components, what they do, the power supply, the hard drive, etc. But we still don't understand the computer - the programming language is a complete mystery. What is on the hard drive?

Finally we crack the hard drive and realize it contains software that isn't written in any language we know. In fact it may not even be designed on the same principals as any computer we understand.

We finally learn the programing language and realize, there are billions of lines of code. Maybe trillions. Now we are trying to read each line, and understand what it does and how it interacts with the whole.

The problem with applying the technological singularity to biology is we have, over and over, vastly underestimated the complexity of the problem we are attempting to resolve. Some of that is just not understanding it was an informational issue to begin with. Molecular biology might be the most complex problem ever identified. It is an emergent intelligence, orders of magnitude more complex than we imagine.

Progress is happening - it has been happening for decades. The COVID19 vaccine is evidence of that progress. I saw they have finally created a program to predict with high accuracy the shape of individual proteins. There are other things like Nurerolink.

And maybe, one day, we get it. But we are trying to crack a 4 billion year old bit of code written (and endlessly re-written) in a language we don't understand that doesn't follow the same computing principals we use. Compared to that, understanding how a star works, or launching rockets to mars, are all pretty simple stuff.

Drew

Bravo!

Kartik Gada

Geoman,

That is brilliant, and it reveals how biology can be viewed under technological principles, just as economics can.

The 4 billion year old+ software is not only still evolving, but the speed at which the most advanced creature evolves continues to accelerate. This trend has never, ever been broken in 4 billion years, even during the greatest extinction events ever, which were 252 million and 66 million years ago respectively.

This is why I believe digital architecture is temporary, and analog will be where computing goes in the future.

Fatcat

@Kartik

Technically speaking the definition of a trend allows for exceptions, as as long there's a trend it is not really broken

Cheers

Joe

As a regular reader of this forum, I think, you should go with better designed but dedicated website where everything future will be discussed. Videos can be linked there too.

It will have much bigger scope.

Attract traffic by regular users linking that site everywhere they can. May be even run some ads initially for traction.

Kartik Gada

Joe,

Do you mean becoming an author on a multi-author blog?

Joe

KG,

Get a future themed domain name. Singularity2050 is not catchy enough.

Your articles and discussions will be the cornerstone. Other serious authors can post too, may be after you/mods vet it. You can reach out to others in your field to contribute.

Let debates happen. Readership will grow organically.

I am sure, this will catch serious attention and you can grow it into a real future themed forum.

Once you gain traction, many of those in power will be forced to take notice, which is currently lacking per my knowledge.

My 0.02c.

Joe


Continuing from my earlier comment.

The comment of the year by Geoman is, undoubtedly, a brilliant article by itself. A lot of discussion can follow from that. So is the 2019 comment of the year.

One way to get more traction for the forum will be to run some ads in online economics forums. student forums, reddit etc.

You can monetize the site by adsense too, to cover the costs.

Kartik Gada

Joe,

In 15 years, traffic has pretty much been the same as what it is now.

The name of the blog is 'The Futurist', but the URL of singularity2050.com is pretty catchy, IMO, due to the self-explanatory nature and round number that makes the URL valuable as the time approaches.

There is already Adsense (in the right-hand side bar). It barely generates enough to cover the $60/yr of Typepad, so the net revenue is zero. I am not really trying to make ad money.

Geoman

Joe - thanks for the compliment. I'm truly flattered.

I'd be happy to contribute a guest article or two to a blog...or not. I have a full time career as a scientific consultant, so I'm not exactly looking for more work or any additional income.

That said, as part of my job, I notice the same problems and constraints cropping up across multiple industries. I hop around quite a bit. Reading Kartik's stuff actually helps me with my job - I'm able to see a little better regarding where the fundamental constraints are. It helps me advise my clients better.

In addition, commenting here is a chance to say things that have been nagging at me for a long time. To work out some demons. I appreciate everyone's indulgence of often overly long comments.


Drew

Hi Geoman, what constraints and problems are you seeing? I know you have written about government not keeping up with change, but I’m interested in hearing more, including other limtations.

Geoman

That would be quite a laundry list.

Regarding constraints and problems, I'm seeing 1) predicting what is going to happen with technology, especially in industries that have not changed much over decades, or have very rigid structures, because the market segment requires it. Utilizes are a major example. 2) predicting future demand for raw materials, like oil, copper, zinc, etc. 3) Communications revolutions, one stepping on another, say starlink competing with fiber optic lines. 4) Sunk capital costs. When do you take the write off? Do you hang on indefinitely? 5) EVs, EVs have profound impacts on utilities, along with cabs, trucking, car companies, battery manufacturers, mining, oil companies, car repair, etc. It's not just a neato new car. It changes the paradigm of multiple industries, with a lot of winners and losers. Every client I work with is talking about them. What will mineral demand be? What happens when battery costs fall below $100/kwh? Will demand explode? Demand for what exactly? Nickel? Lithium? Phosphate? copper? Depends on which cell chemistry wins out. Depends on the political stability of the countries in question. I work with multiple mining companies, and they are all scratching their heads.

6) What business am I even in? A problem I encountered recently - should utilities own EV charging infrastructure? A utility substation would be the ideal location to install charging stalls. Would that interfere with the work the substation is supposed to do? Could it be a profit center for the utility? How will the PUC respond to that? How will the public? Is this a whole new business (service station), or just the same business (selling kwh)?

The other day a friend (an exec at a communications company) and I were speculating about starlink. We came up with this scenario - there are really no issues with automating a ship crossing the Pacific. It's much easier than self driving cars. The challenge is what to do in case of a problem. With starlink you could have a full crew step in in case of emergency. They would check in each day, make sure the ship is on course to it's destination. You no longer have to fully automate the ship - the on-shore crew could step in at any time and run the ship from shore. And the same crew could manage a dozen ships simultaneously. I mean, why not? At port a local pilot team would board the ship, and guide it into the dock. And starlink might cut a very low cost deal - In the middle of the Pacific there are unlikely to be many users. So $100 a month per ship would be a fair charge. The implications are a bit staggering. How much money could be saved by such a system? How will unions react? How will governments regulate this? Can they regulate this? (once the ships are on the high seas...?)

Anyway, getting bogged down in details...

Drew

That makes sense - as in the straight forward version of Moore's law is coming to an end so we can no longer extrapolate when smaller transistors will arrive and go from there (costs dropping by half every two years but in the same computing structure). In this group, we expect similar improvements in computing power going forward, but we don't know from which direction, using what materials and when the increases will occur.

Geoman

Drew - exactly. Or it is Moore's law applying to more and more things?

It occurs to me there is only one problem.

Capital expenditure (capex) is the money an organization spends to buy, maintain, or improve its fixed assets, such as buildings, equipment, or land. A corporation justifies these purchases based on the extended value of the asset. For example, a company will build a factory expected to last 30 years. Each year it writes down the value of that asset until at year 30 value = zero. Or, like with a power plant, the plant is designed to last 30 years..So you take a loan and each year you pay 1/30th the cost of the plant plus interest. It is a very stable and simple system of accounting.

The problem is the singularity is 30 years away. There is no assurance that anything invested in will retain any value over the next 30 years. 30 year amortization schedules increasingly make no sense at all. What capital asset will retain its value that long? It is really impacting capital heavy industries, like power production and large mines, which can last 50 years. How do I know the power plant I build today will have any value in 20-30 years? If it's value evaporates before it is paid off, then I start to get weighed down by debt of useless assets. I have to account for this somehow, but how?

Now, here is the scary part.

Each year we are one year closer to the singularity. Today the problem is large assets with 30 year amortization plans, soon it will be 20 years, then 10, then 5. Shorter amortization plans = smaller and smaller capital assets. What starts of fouling up power plants, and large mines will become roads, bridges, buildings. houses...cars...People won't want to invest the money, believing they will lose it due to rapid technological advancement.

Why build a new highway, when the boring company will transform transportation?

Why buy a new car, when the new Evs will be so much better?

Why build the SLS, when Spacex Starship will be so much better?

Why lay fiber optic cable - why don't we wait and see what low earth robot satellites can do?

Things will become ever more stangnet and confused. Think of it as the Osborne effect, but against society as a whole.

Many industries/governments have never had to face this sort of massive uncertainty before. It's going to be a rough adjustment.

Jenny

"Why buy a new car, when the new Evs will be so much better?"
But this is something the tech industry seems to be dealing with quite okay. People still rush to get the new iPhone and people still buy new computers although they are no longer the best once they arrive home.

Geoman

Jenny - this is something we have adjusted to for a $700 item. It is not something we have adjusted to as a $700 million item.

And I would argue, as nifty as the Iphone12 is, it really isn't that much different than and Iphone 11, or an 8 for that matter. The later retain some value in the face of a new Iphone 12. It also doesn't do much new, it is just the old stuff slightly better. And it has a 3 year amortization, not a 30 year. These are all things we can tolerate.

For the big stuff we are talking about technologies the negate the need for the powerplant, bridge, fiber optic line, mines, or factory. The powerplant is negated by reduced demand due to solar roofs. The bridge is negated by a really cheap tunnel. The fiberoptic line is negated by low earth satellites. The mine is negated by better extractive technologies, and demand swings. The factory is negated by a robotic custom build on demand system.

With FSD and robtaxis, do I even need a car at all? Why spend $40k then?

In your phone example, it would be as if Iphones being replaced by telepathy. And that telepathy is coming...within the next 2 years. And will be free. Would you still buy an Iphone for $700, with monthly payments over 3 years?

That prediction horizon, which is considered a fundamental element of the singularity, impacts finance in important ways. It injects confusion and risk. That leads to hesitation, or higher interest rates. Right now it is just on the edge of our financial perception. Our normal predictive window. But each year it creeps closer, and impacts more decisions.

Kartik Gada

Remember the Upgrade Paradox, where the cost of keeping up with technological upgrades become prohibitive. It is one of the strongest arguments in favor of ATOM-DUES :

https://www.singularity2050.com/2017/04/the-upgrade-paradox.html

Drew

With this framework of mind it would actually make sense that finance, education and medicine would take up so much of the economic space _if_ they were succeeding in their roles of helping us deal with an uncertain future and be around to enjoy it. We could even justify their current share of the economy as searching for a way to carry out those roles successfully, except that so much of what they do is rent-seeking. Counter-intuitively, if they did get better at their roles, there would likely be even more investment in them as society could use the preparation and preservation.

Kartik Gada

Drew,

That is true, but, as you note, education and healthcare are very low productivity.

Hence, they attract ATOM disruption.

But if they were efficient, yes, finance and healthcare would be ever-rising shares of the economy. In healthcare, 50% of all lifetime cost being in the final 6 months of life is unsurprising, and may continue no matter how prosperous society is.

Geoman

One way to save your business model from disruption is to make other business models illegal. I think that is what we are going to see a lot more rent seeking industries.

Look at the fights against Uber and Lyft. Look at the fights over AirBNB. These are just the beginning. Taxis and hotels are small potatoes. The disruptions are only going to get larger, and strike at more powerful interests.

Kartik Gada

Geoman,

These are just the beginning. Taxis and hotels are small potatoes. The disruptions are only going to get larger, and strike at more powerful interests.

Yes. But as I have often said, if 10 groups go to the government to pay for protection money, the government can only cater to 1-2 of them. The rest cannot avail themselves of that.

Plus, there are other countries, in which the disruption can become big first.

Lastly, as tech companies, they achieve high valuations quickly, and thus have the resources to fight back. AirBnB has a higher market cap than most hotel chains combined. Of course, big tech also acts monopolistic, which is another problem.

Drew

Question for the group - do we know what level of resources (in terms of computing resources and human expertise) Deep Mind used for the protein folding problem? Deep Blue beat Kasparov in 1996 and programs on laptops were repeating that level of skill in 2005, but I suspect Deep Mind used even more resources.

Drew

One clarification - I don't mean the human expertise to create the system, but to run it when inputting a new protein to analyze. It took chess experts to create Deep Blue and successor programs but not to run them. I suspect it still takes biology experts to run Deep Mind's protein folding solution.

Geoman

My understanding (and it may be admittedly incomplete) is that Deepmind isn't about raw computational power at all - we have moved beyond that measure of a computer's ability, much like measurements of memory on computers is somewhat irrelevant with cloud storage.

What Deepmind does is machine learning, which is a whole new computational approach. DeepMind took 170,000 known proteins from a public repository of protein sequences and structures. They used between 100 and 200 GPUs (no one seems entirely sure) to calculate protein folds. The program would go from sequence to structure. Bad results are tossed, successful results retained. Training the system on this hardware took "a few weeks", after which the program would take "a matter of days" to converge on the correct answer for each protein structure. It had "taught" itself how to predict protein shapes from the initial sequence.

The old programs like deep blue calculated "moves ahead". That is, look at the chessboard, calculate every possible outcome 5, 10, 50, or 100 moves down the road, and select the outcome most favorable to a win. Obviously seeing further down the road made a win more and more likely, but required more and more computational power.

Deepmind's approach would be to look at previous games that had already been played. Thousands of them. Millions. Billions even. It would look at the final configuration of the board, and back calculate how it got that way from the initial conditions. It would then re-play the games, trying to predict the outcome (the shape of the pieces on the board resulting in victory). Using this it would slowly develop an algorithm with a higher and higher level of success. Weeks of "training" would result in an algorithm that was highly likely to win each time.

Now during a real game, it wouldn't calculate moves ahead, it would say "in such and such a situation, move X is most likely to cause the final configuration of the board to match one where the computer wins." It is unlikely to ever be 100% successful, but it will be fast,and use little computational power. And it can get better over time, with success rates in the 90 percentile.

In a way it is thinking, much like we do. We don't win at chess, or Go by calculating millions of moves ahead, instead we develop, based on experience, a set of internal rules that govern our choices. Sometimes we are not even aware we are doing this.

To become an expert in anything requires 10,000 hours of practice? Well, what IS that 10,000 hours doing exactly? It is your brain slowly generating an algorithm that has a very high potential for success.

Look at basketball. there are players known for their "downcourt vision" basically they are so good handling the ball that they don't have to think about that. their mind is occupied with looking downcourt and seeing the positions of the players and predicting likely final configurations that result in a basket being scored. That takes thousands of hours of practice to achieve. Some players never get it. It's why guys like John Stockton excelled - he wasn't the biggest player, the strongest, but his court vision was phenomenal. His brain contained algorithms that were highly successful at predicting positive outcomes.

Anyway, that is just my understanding.

Geoman

To answer your second question (which arrived while I was composing the first answer), I don't think it takes biology experts to run Deepthink at all. A biologist might be asked to posit a starting algorithm for predicting a protein structure based on protein sequence. Just a guess at what the formula may look like, the variables involved. The computers do the rest.

Imagine 200 GPUs. The input is protein sequence. Each runs the sequence through the algorithm. The output is the structure. Each compares the calculated structure to the real structure - it is not the same. Now comes the cool bit, each tweaks the algorithm slightly, in some random way. The results are compared again to the answer. The closest answer is input back into the 200 GPUs, and each runs a tweak again. Tweak run, improve. Over and over. With 200 GPUs doing this thousands of times a second. With the closet correct result being re-input into the system. In a matter of weeks (which good gravy, how many tweaks could that be? Quadrillions?) the algorithm gets closer and closer to perfect prediction. It is checked against the 17,000 proteins we know the structure of.

Now you can input a new protein sequence to any GPU with the correct algorithm with 95% certainty it will predict the right protein structure. This is no longer a massive computational project - the right computer with the right code can do this.

And, as 17,000 starting proteins becomes 100,000, and as the GPUs run for months and years as opposed to weeks, the predictions will get increasingly accurate. It will never be perfect, but it will likely be very very good.

What we are trying to do is bridge the gap between two things we know, the sequence, and the structure. Once we know what formula goes between the two, we can then back calculate the missing information. Start with a desired structure and back calculate the sequences, or start with the sequence, and determine the structure.

The value proposition is that sequences are EASY to determine, structure has always been a huge pain in the ass. In my callow youth I used to do x-ray crystallography of minerals. What a pain! We'd know the chemical formula, but not the shape of the mineral crystals, which was important, because the shape had a lot to do with physical properties. Just look at graphite and diamond. Problem is you need a high power x-ray source. And with organics proteins the problem is much harder to solve - you fry the molecule before you can get enough diffraction to get an image. I felt sorry for us, but those poor molecular bio saps....

Anyway DNA was discovered in the 60s. The 1860s I mean. By 1881 we identified nucleic acid and provided its present chemical name, deoxyribonucleic acid (DNA). And we isolated the five nucleotide bases. It took until the 1953 to figure out the structure of those building blocks, the actual shape of the molecule. Even today, determining the shape of proteins experimentally is slow going. Not 70 years slow, but...months and years man effort per protein. Yet we can figure out the sequence, often in seconds.

So for the most part, this will be used to take sequences and predict structures. "Why does this protein do that, while this other protein does something different? Must be the shape...." But in theory, we might want to have a protein of a certain shape, say a medicine, and we can predict the sequence that will generate it. This is the key to molecular medicine and designer drugs.

Drew

Thanks for the great answer. I'm trying to reconcile how we can have lots of evidence of declining productivity in scientific research (cost to develop drugs, cost to move to the next smaller transistor size) with the much better tools we have now. It's a version of the old economic proverb about seeing computers everywhere except in statistics. I guess I will go with that futurological proverb that change takes longer than we expect, but when it comes it has a greater impact than we expect.
It may be 20+ years from now when the current graduate students are running labs of their own for the full impact will be felt.

AL Villalobos

Slightly off topic - Back in September, Mr. Garda , talking about poor overall government efficiency said:
"Also, there is immense upside for a small, high-tech country to decide that it wants a super-advanced, highly efficient government. Capital will flow in, and competitive pressure will emerge in the sovereign governance sector."
Would anyone care to throw out a list of the "Top 5" countries to watch?
I'd like to be ahead of the curve - for once!

Kartik Gada

AL,

See Chapter 10 of the ATOM publication.

Now, being the most suitable does not mean it will be the first to do it. Singapore is conservative, and always wants to wait until someone else does something first. Hence, even if Singapore is much more suitable than Estonia, Estonia is more adventurous and is more likely to do it before the much-wealthier and more high-tech Singapore.

Joe

Geoman,

And that telepathy is coming...within the next 2 years. And will be free.

Care to expand on these? Sources to read?

Geoman

Joe,

I meant that as an example of how the Osborne effect works. The idea being you are aware that, within a relatively short period of time, a much better product will utterly disrupt the product you are being asked to pay $700 for. You can read up on the Osborne effect - sad story about a technologically advanced (perhaps even the best) computer company shot themselves in the foot, and destroyed their brand. Us old guys remember it well.

And the bigger the risk, the more powerful the effect. It is one thing to risk $500, but to risk $500 million. Especially if you are taking a loan for the $500 million.

Telepathy is highly likely to be scientifically impossible. Maybe we could use neuro link coupled with cell phone technology to approximate such a thing...but it is at least a decade away from even experiments. And I very much doubt anyone will be willing to stick wires in their brain so they can make faster cell phone texts. The real use will be remote controlling machinery.

Kartik Gada

Folks,

I am coming to the conclusion that migrating this content over to a YouTube channel as the primary output, will probably be best. In fact, perhaps I should have thought of that a decade ago! Or at least five years ago when the ATOM was published.

But all is not lost. I already have enough ideas for the first 100 videos (10-20 mins each), on account of 15 years of archives here. YouTube has little to nothing in the way of Futurism, particularly as meshes with real-world business and finance, and other big-name Futurists, such as Ray Kurzweil, have had little to no output in the last decade.

As the education bubble is truly being disrupted as we speak, and my teaching at Stanford has been on hold for a long time due to Covid, the time is now. This can synergize more seamlessly with my normal professional activities as well, reinforcing everything.

More details to follow.

Geoman

http://www.rationaloptimist.com/blog/mrna-vaccines-could-revolutionise-medicine/

Nice non-technical discussion on why the COVID19 vaccine is revolutionary. After COVID19 is dealt with, expect vaccine development and deployment to take a couple of months...for just about anything potentially including cancer and allergies. It really will change the world.

https://www.bloombergquint.com/onweb/some-u-s-colleges-cut-tuition-ending-relentless-price-spiral

The slowly bubble bursting for higher education.

Congrats on the move to youtube!

Are you going with the talking head approach, or some form of animated presentation with voiceover? I prefer the later myself - graphs, charts, pictures...help drive the message.

Kartik Gada

Geoman,

Combination of talking-head with the type of graphs and images that I post here. Some talking-head time is necessary as I want inbound leads on consulting, speakerships, TV spots, and corporate directorships.

I don't have major animation skills yet, but we will see. Many videos will link back here, in the form of '12 years ago, I predicted this. Here is how that turned out'.

Tying real-time financial/economic/technological news to videos will also be easier under the higher frequency of video output. That could generate real money too, when affiliate marketing factors in.

This current blog makes no money (it is at breakeven, since Typepad costs $60/yr, while Google ads have been $60 in the last year). Peanuts all around. I always saw it as my hobby.

But after investigating multiple YouTube channels, this one will have no trouble getting to $5000/month before too long. Not that I need $5000/month, but the virtuous cycle of inspiration to keep doing more is very immediate and obvious.

More importantly, it synergizes with more of my actual professional activities much better, such none of my professional peers have any YouTube channel or other significant professional social media presence beyond LinkedIn. What I teach at Stanford should also be available widely, rather than to just the super-select few who have the extraordinary privilege of going to *Stanford University* in person.

Drew

Good luck and I agree that it should drive more interest as well as gigs since people will see how you present. As a person with a strong preference for transcripts over videos, if you run the videos through a voice dictation system, I volunteer to edit the transcripts to read correctly - I will guess three years until computers do that better than me ;)

Kartik Gada

Thanks, Drew! That is a good idea.

In fact, some of the transcripts can still be posted over here on the text blog.

Kartik Gada

Geoman,

https://www.bloombergquint.com/onweb/some-u-s-colleges-cut-tuition-ending-relentless-price-spiral

The slowly bubble bursting for higher education.

Wow, Oberlin knocking $10,000 off of annual tuition! That is a big indicator, as that is a high-ranked school in its category.

The fact that they had to pay a $44M settlement to a bakery they abused might have something to do with it.

But I am still hoping to see a mass layoff of administrators. The change is not structurally deep until that happens.

There is no reason universities cannot return to the cost structure of the 1980s (or earlier), as measured by the tuition per year relative to the starting salary of its grads.

Geoman

Oberlin is still fighting that lawsuit, despite having almost zero chance of prevailing. And the ongoing fight has only served to increase their liabilities, which started at a measly $250k. It's quite comical, yet illustrates the obscene misplaced priorities of so many universities. How did any of this serve the students?

I have a way for colleges to 1) save boatloads of money 2) 100% comply with any measure of fairness.

1) Incoming potential students take a standardized test. The results are without any personal identification - students are assigned a random ID code. No information is collected on the students at all.

2) Tier 1: Scores > a certain number receive automatic admission offers. Tier 2: Above the cut off, but below automatic admission.
Goes to a lottery system. Admission is offered based on the lottery results. Tier 3: Scores are below a cut off number and are automatically denied admission. Each school can set their tiers each year. Maybe the top tier is simply the top 20% of applicants, and the lowest tier is the bottom 20% of applicants.

3) Once admission is offered the names of the students are revealed. The school can, at that time, offer scholarships to encourage attendance.

4) If a student declines admission, the lottery spits out another name. Keep going till you run out of students in the middle tier, or all your slots are filled.

A perfectly fair system free of all preferences for admission. It can be used for both American and foreign students.

If a school wishes to increase their % of a certain racial minority, they must 1) tutor those students to improve their test performance and b) encourage them to apply. Universities are free to provide additional education to any high schoolers interested in college. Ideally, a university might "sponsor" a local high school in a poor part of town, offering additional guidance counseling and tutoring to get more minority applicants.

Imagine how much of the admissions staff at a university could be let go under this system. No essays to read, no interviews, no boards. Everything fair and equitable. Very little oversite necessary. It runs almost automatically. It ensures the best and brightest get placement. It encourages the university to improve the quality of their incoming students. It focuses high schools on teaching kids what they need to know to pass the test.

My daughter recently passed her exam to be an RN. Fascinating test. It asks a question. You answer. If you are correct, it asks a harder question. If you are wrong it asks an easier one. It keeps asking questions till it determines your overall rating. You may get 100 questions, you may get 500. It depends. No two tests are alike. yet the results are highly accurate predictors of your education and abilities. And the only way to beat the test is to really know the materials. A test like that, focused not so much on knowledge, but the logic and thinking skills necessary to do well in a college setting, might be appropriate.

Kartik Gada

Geoman,

All true. But of course, the Universities themselves are filled with people who have exactly the opposite value system of what is needed to do what you describe.

The cost bloat relative to 1980 is just insane.

Geoman

Any school that has an admission policy that relies on the judgement of an individual, or small group of individuals, in inherently unfair. No matter how enlightened the individual or group. It is subject to corruption, racism, and cheating. The only purpose of complicating the admission decision is to allow individuals to unfairly influence the process.

It is no coincidence that in 1978 the student loan act passed congress.

Kartik Gada

Geoman,

A test like that, focused not so much on knowledge, but the logic and thinking skills necessary to do well in a college setting, might be appropriate.

I would say a test like that should be used by employers, giving 18 year olds the chance to be hired straight from there, and bypass college altogether if they want.

No matter how enlightened the individual or group. It is subject to corruption, racism, and cheating.

As far as admissions, this is something that could very easily be algorithmic.

SAT scores.
GPA relative to others from the same high school over the last 20 years, indexed for difficulty of courses taken.
Essays : The algorithm can rank it easily relative to the millions of other essays.
Extra-curriculars : There are thousands who have done the same thing. This too can be algorithmic.

While some might say that the 'subjective' element is taken out, the fact that admission decision to 10 (or whatever maximum number of schools the student can apply to via algorithm) universities can be instantaneous (and free), is an advantage.

Of course, the failure of universities to take these steps, main for reasons of a faulty value system across administrators in Higher Ed, is why online alternatives are really taking over.

Drew

The university.is selecting for the most prestigious peer group, not academic talent, but even that could be automated. Hmmmn, maybe I have to step up and create my match up for pods after all.

Kartik Gada

Here are the first ~180 YouTube videos for which I already have the entire content in my head, just from rattling off some preliminary brainstorming on a blank page. Some titles are self-explanatory to readers here, while others may sound unusual and unrelated, but I assure you they tie back to the broader theme. Part of what will help traction is addition of topics relevant to the tech industry, careers in tech, and personal finance. The mix between high-concept topics and more mainstream videos will be seamless.

These are just the first ~180 that I already have in my head in their entirety, and do not include videos on relevant hot new news items in tech, finance, economics, etc.

At this rate, it seems like it will take 300+ videos for me to catch production up to ideas I already have.

January will be spent in the learning curve of video production and editing, and buying equipment. February will be the first shooting, as daylight availability will be better by that time.

I hope I can become skilled enough at editing that output reaches 25% of total production time (i.e. 1 hour to produce a 15 minute video, including filming, editing, uploading, SEO, etc.). At first, I assume it is 3 hours.

No expectation of significant traffic until 100+ videos are posted, which is usually where the inflection point occurs. This will take 300+ hours of time (average 15 minutes per video). I hope videos 101-200 take only 100 hours of time, and so forth.


__________________________________________


Why is a Complex Tax Code Ruinous for the Economy?
Why Technological Disruption is Good
Why the Establishment 'Experts' Usually Turn Out Wrong
How government must manage tech monopolies in the high-tech age
Why the Technology Sector Makes So Much Money
The Accelerating Rate of Change
The Accelerating Rate of Economic Progress
Who is a Futurist, and who isn't?
How much of Tech success is luck?
What I teach at Stanford University
Don't move to Silicon Valley Unless….
Most World Poverty Will be Gone by 2032
Why is India Still Poverty Stricken?
Don't Overrate India's Technological Talent Pool
College, As Exists Now, Leaves You Underprepared for the New Economy
Technology and Organized Crime
The Biggest Tech Failures : $Billions Lost
Everyone Must Learn to Code
Why the Top Lawyers Make So Much Money
Economic Progress : More Skillsets Monetized
Executive Compensation : Everything You Wanted to Know, But Were Afraid to Ask
Sartorial Excellence : Why Dressing Well Still Wins, Especially in Tech
The Anatomy of a Technology Company Acquisition
Technology is Borderless, which makes the disruption of Government inevitable
The biggest technological disruption of the Next 20 Years
Technology Always Finds a Way
Economists : No Other Group Knows Less About Their Own Field
Let the Economy Breathe from BOTH Lungs, for a Change
Ever-faster Consumer Technology Adoption
You deserve Federal Reserve Money!
Overhyped Technology : Autonomous Vehicles
Overhyped Technology : Quantum Computing
Celestia : Free Education
Will there be another recession?
YouTube is an exceptional metabolizer of Federal Reserve Monetary Creation
How the ultrawealthy manage to not pay income tax
Due for Disruption : Legal Services and Legal Fees
Due for Disruption : Healthcare costs
Don't be an Angel Investor in Technology
Does Real Estate have a Good Return?
Where have all the Futurists gone?
There will be no hyperinflation
Things people in the future will find odd about our time (a Series)
My First Exponential Technology
Automation is Your Friend
FinTech : A Primer
Tax Automation, not Humans
The One Difference Between an Advanced Country and a Dysfunctional One
Can Government be Automated, and Made Low-Cost
The Dow Industrial Average vs. S&P500
Comparison of Equity ETFs
Why March 15, 2020 was the Netscape Moment in Economics
The history of Quantitative Easing
Why the Fed Funds Rate will, and should, be nearly zero forever
Why Monetary Creation (Quantitative Easing) has to be Permanent and Exponential
How the Stock Market has Trained the Federal Reserve into Compliance with the New Economy
Why you can never ban a technology
YouTuber : Tell us your story of success
Why Indians Have a Low Assimilation Rate in the West
The Capitalism of YouTube
Your Purpose in the High Tech Age
The Paradox of Capitalism : Perverse Profit-Seeking
The Technological Singularity
Artificial Intelligence and FinTech Reports
Artificial Intelligence : Intro
The Center of Gravity of the World Economy
The Importance of Starlink
Distributed Ledger/Blockchain Intro
Cryptocurrency - A Technology?
Cousins, Part 1 and 2
The Education Disruption
My 23andMe
What does a Board of Directors do?
The QSBS Deduction for tech investors
Qualifications to be a Board of Directors member.
Do you need an MBA in the high-tech age?
What does a CFO do?
Why the CFO job causes so much confusion
Why Project Management is the best mid-level Tech job
What does a CTO do?
Technology both Assists and Disrupts Lawyers
How Lawyers can get into the tech industry
Why a big Tech Company needs a Chief Economist with no Economics Degree
Which Technology Headlines to Ignore
The fundamentals of getting your tech company acquired
The Truth About Longevity Treatments
Are you ready to live to be 100?
SETI and the Singularity
The Technological Purity Score
Why Private Equity has become such a large factor in the Tech Sector
Myths about the Midas Touch of Venture Capitalists
Tax Simplification in One fell Swoop
How to become a Senior Executive at a Tech Company
Tech Unicorn Showcase
Iownit Platform
Banks are Obsolete
Why Production of New Knowledge is the Highest Possible Pursuit
The Evolution of Intelligence
Major Extinction Events, and the Megatrend of Intelligence
UN HDI
Yardeni World QE
The Cambrian Explosion of Money
BRIC'? Fact and Fiction about Emerging Markets
Why India is not a high-growth economy
The Law of Finite Petrotyranny
Why Corporate Careerists Need a YouTube channel
When Low-Tech Converts to High-Tech
E-Commerce is a Bigger Disruptor Than You Think
Domain Name Investing : Does This Still Work?
The Individual ATOM
How to 'ATOMize' your YouTube channel for Futureproofing
ATOM Stories
Even Andrew Yang was 'Faster Horses'
The only real path to UBI (Universal Basic Income)
World Poverty Reduction Trend
Hopelessness
The Real Inequality
ICE Cars vs. EVs
India Factoids
Indian-Americans
Why Automation Creates Jobs
Traffic by Decade
Technology has changed behavior : TV Series Seinfeld
Technology has changed behavior : TV Series Friends
Movie Assessment Soylent Green
Movie Assessment Star Trek : First Contact
Movie Assessment 2001 : A Space Odyssey
Movie Assessment 2010 : The Year we Make Contact
The Upgrade Paradox
The Impact of Computing
Why you must Contribute to your 401K
To Help Society the Most, Do What You Do Best
Self-Driving Cars
I have Predicted Stock Market Tops and Bottoms Many Times
Don't own Gold
Don't pick stocks
Don't let politics fill your mind
Review of Singularity University
Employers have to Step Up to hire people without College Degrees
My Innovation Prize : The Uplift Prize
Where is 3D Printing
The Case for Optimism
Graham Stephan Did the Right Thing by Not Going to College
Job Searches Worsened by Technology
How I Became a Leading Artificial Intelligence Expert
Why Deflation is far worse than Inflation
Book Review : How We Got to Now
Book Review : Telecosm
On Carl Sagan
On George Gilder
On Ray Kurzweil
On Harry S Dent, every prediction wrong
Why PhD Economists know nothing about the real Economy
Technological Progress of Video Games
Equity Prices, not GDP, are the real measure of Economic Progress
Why Tech Unicorns Want to Avoid Going Public
The Accelerating TechnOnomic Medium : Executive Summary
ATOM AotM Jan-17
ATOM AotM Feb-17
ATOM AotM Mar-17
ATOM AotM Apr-17
ATOM AotM May-17
ATOM AotM Jul-17
ATOM AotM Aug-17
ATOM AotM Sep-17
ATOM AotM Oct-17
ATOM AotM Nov-17
ATOM AotM Jan-18
ATOM AotM Feb-18
ATOM AotM May-18
ATOM AotM Oct-18
ATOM AotM Dec-18
ATOM AotM Feb-19
ATOM AotM Apr-19
ATOM AotM Nov-19
ATOM AotM Feb-20
ATOM AotM Jun-20
ATOM AotM Aug-20
ATOM AotM Nov-20
Politics is a Circle
Does Immigration Suppress Wages?

Drew

Hi KG, I do not have video editing or animation skills, but if I can assist beyond editing transcripts, please let me know.

Drew

PS - I am a project manager, but applying that to a one person operation is often overkill

Geoman

That is...some list. But you've been doing this blog sine 2006, and a large number of your topics are "evergreen". In fact that is what you have strived for - finding topics to discuss that are not transitory, but highlight long term trends in society.

If you ever run out of ideas (and I doubt you will), there are always people here willing to suggest new topics. And I must say, the bar for this sort of thing is very low. I'm shocked at how often I watch a Youtube video that I think might discuss something interesting, and the results are simplistic, not well thought out, obvious, badly presented or organized, or just boring. If I have one suggestion, it would be this - stuff is more interesting than finance. That is to say, people love technology videos. Videos on corporate tax rates....not so much.

I wish you the best of luck with this new endeavor.

Kartik Gada

Hello Drew,

Yes, certainly. I said that above on Dec 28, at 12:24 PM at well.

We'll have to see how that can work. There will be images, etc. in in videos that won't be captured in the transcripts.


Hello Geoman,

Thanks! This website will still exist, but with few new postings (not that it is frequent now).

But content for at least 300 videos is not going to be a problem at all.

I'm shocked at how often I watch a Youtube video that I think might discuss something interesting, and the results are simplistic, not well thought out, obvious, badly presented or organized, or just boring.

Yes, which is exactly what I noticed, and how I could outcompete a lot of successful channels, even as a late arrival. Channels on personal finance are very rudimentary. But the traffic and ad revenue is often still impressive for such channels.

Plus, from the other end, none of my professional peers, and no other Futurist, has any channel at all, successful or otherwise. My extensive experience in public speaking, and the Stanford connection, will help here.

Kartik Gada

Geoman,

If I have one suggestion, it would be this - stuff is more interesting than finance.

Most will be futurism, or items related the the technology industry and technology careers.

Videos on corporate tax rates....not so much.

Generic videos, no. But a video that highlights and explains a certain tax shelter, could do very well, as people are always scouring the Internet for tax shelter information, and extremely few tax lawyers even have good websites or good advertising. Ultimately, it all fits into the topic of 'money', that people are always trying to learn more about.

The orthogonal points about the inability to tax AI, and the ATOM points, all get a synergistic boost through the SEO.

Philipp Stelzel

YouTube is a great platform for attracting more people. Still, your blog is a wonderful asset: Use the existing traffic and post your new videos here as well. If you have several hundred visitors per day, that should give your channel a boost.

You could also embed videos into existing posts. Personally I prefer text to video because you can cover a topic much more deeply, therefore I am hoping that you will continue blogging once in a while.

Revenue: Ads are the most ineffective way to make money on the internet. I would rather recommend books about singularity or technology in the sidebar. You could use Amazon affiliate links and earn commission for every sale. That should work much better.

Kartik Gada

Hello Phillip,

Thanks. At the moment, the ad revenue is very little.

But if the affiliate links are just obscure books that few people buy, would that actually generate anything?

Yes, the blog posts will be around the videos, assisting the traffic of the videos. Since the videos will be far more frequent than blog posts, we shall see how that can work.

Not much will happen until at least 100 videos are posted, since the effect is very exponential. All those videos are already in my head.

But I will have to spend $1000+ on Equipment, and perhaps $2000+ on editing those first 100 videos. That is very little relative to starting up almost any other business (and there is no chance I won't make at least that much back), but that is the outlay for now.

I am happy if the channel eventually makes just $5000/month. What it could do for my normal professional activities, on top of establish me in as one of the topmost futurists (i.e. when viewed by people 40 years hence, as well as post-Singularity), is the main goal.

We are actually in a post-futurist era at the moment, ironically enough, as all the big-name futurists of 15-20 years ago have gone quiet (perhaps due to old age), and there is a vacuum. I can fill the vacuum quickly.

A place like Singularity University has fallen into disrepute, and a channel like mine can provide all that content and much more, for free (Sing U charges $15,000 for a one-week onsite program, which is about as far away from ATOM-type thinking about business structure as one can get).

Fatcat

Kartik,
depending on what you currently have you might start with much less investment. There is a bunch of open source video eidtiing software. It takes time to be proficient but the basic stuff can be learned in a few hours. Virtually all modern smartphones (at least mid-tier ones) have decent cameras with the right lighs. If you have a decent bluetooth micrphone than all you need is a PC with enough dis space, a smart phone , background , good lights, and a phone/camera mount. Good lights might actually be the trickiest part.

Kartik Gada

Fatcat,

True, but in researching a lot of videos that compare cameras, I might have to bite the bullet on a $650 Canon M50. The marginal quality improvement matters a lot.

Lights are actually cheap, as there are a ton of hacks around the need to buy $1000 lights (hacks with LED strips, AL foil, etc.), plus there is good daylight here in the Bay Area after Feb 15 or so, with almost every day guaranteed to be sunny. I already have Tripods and Microphones. A $30 SD Card and $100 USB external storage unit are also needed.

Video-editing : The main thing is the learning curve, and the time I might spend. I will do some myself, while shopping for editors in low-wage countries on Upwork and Fiver.

Assuming ~20 mins of raw footage for a completed 12-15 min video, paying $1/min of raw footage is ok (~$20 per video). The key is to get to 100 videos several months sooner than I otherwise would if I edited them myself.

There is the counterargument that editing videos oneself doubles as time that sows the seeds of new ideas, as well as forces scrutiny for more improvement. So I might edit half and outsource half of them, or something.

I am striving for a solid professional standard of videos here. Going $3000+ in the red at first is no problem at all.

Geoman

Go big or go home.

I agree you should invest what it takes to make things look as professional as possible right out of the gate. Your ideas are worth sharing - don't want people getting hung up on silly stuff like poor video quality, etc.

On another note - I believe I once ranted on how all the technologies developed for oil and gas fracking would eventually be utilized for geothermal power. I just ran across this the other day.

https://deepcorp.ca/#:~:text=DEEP%20is%20a%20Saskatchewan%20privately%20held%20corporation%20with,%28or%20earth%20heat%29%20power%20facility%20in%20southern%20Saskatchewan.

I think you could make this system even cheaper and better using advanced thermoelectrics. And you could adapt this system for use in depleted oil fields - meaning your installation costs would be almost nil.

Fatcat

One disadvantage of vlogging is that the expectations are for high quality video and audio . We are spoiled by tv shows which can easily cost 1 million per episode. In the last few years the advertisement networks pushed "pivot to video". Which causes a 20 minutes youtube video which has information content of 3 paragraphs of text.
It is harder to search and skip videos and easier to inject adds.

If your goal is not to be a youtube influencer (a new word);then don't fall for the trap of creating video bloat

Drew

So, while I'm glad that Bezos and Musk have billions to spend on space colonization (especially since space colonization's economic rationale is essentially that it would be cool to do) - I'm seeing the stock prices as more of a sign of the deflation portion of ATOM and lack of other opportunities rather than as a sign of economic strength. Perhaps as a measure of the best version of deflation?

Kartik Gada

Drew,

Absolutely correct. There are too few destinations for so much capital. Even cryptocurrencies have doubled in the last 6 weeks. Meanwhile, the ineptitude of the government in transmitting stimulus funds leads to a lot of tech startups going under.

ATOM-DUES would have been vastly better, as it is a diffuse transmission of money, and has the greatest impact on the neediest people. Everyone from bootstrapped tech startups to unemployed restaurant workers would be in a better place today.

Geoman

We know, not theorize, but KNOW, that the marketplace acts as an emergent superintelligence many times smarter than any human, or group of humans, can ever hope to be. Why do index funds beat stock pickers over time, consistently? Because stock pickers rely on a single human intelligence, while the index fund relies on the emergent intelligence of the marketplace. You can see and measure the differential. It is persistent and universal.

A diffuse injection of funds into an economy will direct the funds precisely to support the businesses worth saving, and away from those that are lost causes. Only the best will be saved, and more will be saved, than any other alternative. No politician can direct the money more precisely or efficiently.

Verify your Comment

Previewing your Comment

This is only a preview. Your comment has not yet been posted.

Working...
Your comment could not be posted. Error type:
Your comment has been posted. Post another comment

The letters and numbers you entered did not match the image. Please try again.

As a final step before posting your comment, enter the letters and numbers you see in the image below. This prevents automated programs from posting comments.

Having trouble reading this image? View an alternate.

Working...

Post a comment

Your Information

(Name and email address are required. Email address will not be displayed with the comment.)