Ask HN: Why do so few people work on important problems?
13 comments
Because important problems (to me, like hunger, health, shelter, etc) tend to cost money rather than make money. To me, most ventures nowadays are about making money more than solving big problems, because the problems have mostly gotten worse in my lifetime.
Please allow me to rephrase..
I do what pays me the most to do. In 2020 I was running 4 interviews at the same time for 4 different companies: 1) medical 2) bank 3) IT services 4) engineering. All 4 were world-class/mega-big corporations. I got an offer from each. I went to the Bank. My heart was dead-set on the medical.. for real.. I was 'praying' they would make the best offer, or a second-best-by-5%-lower. The bank offered 20% higher than everyone, so Bank it is!
Idealism is good, but money is juuuust a little bit better (within the boundaries of morality and ethics)(I wouldn't trade heroin for 10x, but I would go to a big-pharma for 20% more).
I do what pays me the most to do. In 2020 I was running 4 interviews at the same time for 4 different companies: 1) medical 2) bank 3) IT services 4) engineering. All 4 were world-class/mega-big corporations. I got an offer from each. I went to the Bank. My heart was dead-set on the medical.. for real.. I was 'praying' they would make the best offer, or a second-best-by-5%-lower. The bank offered 20% higher than everyone, so Bank it is!
Idealism is good, but money is juuuust a little bit better (within the boundaries of morality and ethics)(I wouldn't trade heroin for 10x, but I would go to a big-pharma for 20% more).
And it is the choice of the money, over solving the problems that chasing money creates, that perpetuates our global disorder.
We all have different definitions of important problems.
For me, Secure General Purpose computation is one of the most important problems, but most people don't seem to even realize the scope/nature of the problem. You've got an entire industry focused on the wrong problems, and thus offering the wrong solutions.
The problem is one of a huge missing feature, the ability to spend a chunk of your compute resources (CPU time, files, etc) on a problem in a safe and secure manner.
Because of this, much effort is being put on "safe" languages, "safe" programming techniques, teaching users how to be "safe", and updating/patching everything, every day, on every device in the world.
At it isn't working. That approach can't work.
Thus, computing isn't something you can trust. You're left in a world of potential land mines everywhere, metaphorically. This then makes any play outside of the cleared zones (FAANG run sites) unsafe. So effectively, general purpose use of the internet is no-go to 99.9% of the public.
This gives the gatekeepers almost unchecked power to control discourse, and subvert Democracy.
It can be fixed, the solutions have been known since the 1970s. There's no general interest in this, nor even a perception that the problems can be solved. We've generally convinced ourselves that computers require the services/subscriptions from experts to "keep them safe"
Would you want a locksmith continuously updating your locks on your doors? Why put up with one on your computer?
For me, Secure General Purpose computation is one of the most important problems, but most people don't seem to even realize the scope/nature of the problem. You've got an entire industry focused on the wrong problems, and thus offering the wrong solutions.
The problem is one of a huge missing feature, the ability to spend a chunk of your compute resources (CPU time, files, etc) on a problem in a safe and secure manner.
Because of this, much effort is being put on "safe" languages, "safe" programming techniques, teaching users how to be "safe", and updating/patching everything, every day, on every device in the world.
At it isn't working. That approach can't work.
Thus, computing isn't something you can trust. You're left in a world of potential land mines everywhere, metaphorically. This then makes any play outside of the cleared zones (FAANG run sites) unsafe. So effectively, general purpose use of the internet is no-go to 99.9% of the public.
This gives the gatekeepers almost unchecked power to control discourse, and subvert Democracy.
It can be fixed, the solutions have been known since the 1970s. There's no general interest in this, nor even a perception that the problems can be solved. We've generally convinced ourselves that computers require the services/subscriptions from experts to "keep them safe"
Would you want a locksmith continuously updating your locks on your doors? Why put up with one on your computer?
what solutions do you mean?
The Capability Object Model, AKA, Capability Based Security. It was invented in response to issues discovered during the Viet Nam conflict by the US Military. The Bell-LaPadula model is an example of a policy model that can keep a computer secure while keeping it usable.
The permission flags mess on your smartphone is NOT capability based security. Neither is AppArmour, or SeLinux. Linux is completely incompatible, as are Windows, MacOS, etc.
Interestingly, it's not much change in code for an application to be ported. You take out code for calling file selection dialogs, and file opens, and replace then with a call for a powerbox to select a capability, and use that in place of the file handles, and you're done.
It's more like cash in your wallet... you can easily take out $5 to pay for something, limiting the side effects of the transaction to $5. It's not possible to run code (without gymnastics) and only give it file X on most operating systems.
WASM is as close as we've gotten in a while since they gave up on Multics. Genode is coming, but they seem to have been distracted by smartphones.
The permission flags mess on your smartphone is NOT capability based security. Neither is AppArmour, or SeLinux. Linux is completely incompatible, as are Windows, MacOS, etc.
Interestingly, it's not much change in code for an application to be ported. You take out code for calling file selection dialogs, and file opens, and replace then with a call for a powerbox to select a capability, and use that in place of the file handles, and you're done.
It's more like cash in your wallet... you can easily take out $5 to pay for something, limiting the side effects of the transaction to $5. It's not possible to run code (without gymnastics) and only give it file X on most operating systems.
WASM is as close as we've gotten in a while since they gave up on Multics. Genode is coming, but they seem to have been distracted by smartphones.
If you define “important problems” as problems that directly affect human welfare (hunger, access to important information, human health, economic opportunity/pursuit of happiness, public safety, military security), each of these have tens to hundreds of millions of people working on them every day. They are called farmers, doctors, nurses, educators, installation and maintenance personnel, sanitation workers, soldiers, public health professionals, caregivers, etc. Actual workers.
If you mean “visionaries” selling self-serving snake oil solutions to what often prove to be non-problems or insolvable problems, there are rather fewer of these than workers directly implementing daily solutions. Thank goodness for that.
If the visionaries had many years of practical hands-on experience solving day-to-day problems before proposing their broader schemes, more would bear fruit and more visionaries could be funded. But few want to do that preparatory dirty work. And those that have that experience are more valuable to society as managers of the daily work and generally lose their enthusiasm for the endless posturing that goes with high-stakes political crapshoots masquerading as “solving important problems”.
If you mean “visionaries” selling self-serving snake oil solutions to what often prove to be non-problems or insolvable problems, there are rather fewer of these than workers directly implementing daily solutions. Thank goodness for that.
If the visionaries had many years of practical hands-on experience solving day-to-day problems before proposing their broader schemes, more would bear fruit and more visionaries could be funded. But few want to do that preparatory dirty work. And those that have that experience are more valuable to society as managers of the daily work and generally lose their enthusiasm for the endless posturing that goes with high-stakes political crapshoots masquerading as “solving important problems”.
Let's run through YC's list: https://www.ycombinator.com/rfs
Most of these are bloody expensive - carbon removal, cell therapy, medical devices, longevity, transportation and housing, education, etc, etc.
Government 2.0 is just well, really difficult.
Which of these are cheap? AI? Biohacking? Supporting creators?
The last one is fairly doable, but there's probably people tackling it, including the people who host the sites these creators are on. It's also arguably the least important one on there.
AI and biohacking have a huge skill barrier. But both of these are very well funded because of that moat. Most of us can do some AI wrapper, but the high funding means that it moves fast enough that most of these wrappers are being replaced by official OpenAI stuff.
So you're left with the problems that are too difficult and expensive to tackle. Some people, like Gates and Musk try to take them, but there's only so many such individuals. There's a lot to say about billionaires funding these projects, but I'll leave that to someone else.
Most of these are bloody expensive - carbon removal, cell therapy, medical devices, longevity, transportation and housing, education, etc, etc.
Government 2.0 is just well, really difficult.
Which of these are cheap? AI? Biohacking? Supporting creators?
The last one is fairly doable, but there's probably people tackling it, including the people who host the sites these creators are on. It's also arguably the least important one on there.
AI and biohacking have a huge skill barrier. But both of these are very well funded because of that moat. Most of us can do some AI wrapper, but the high funding means that it moves fast enough that most of these wrappers are being replaced by official OpenAI stuff.
So you're left with the problems that are too difficult and expensive to tackle. Some people, like Gates and Musk try to take them, but there's only so many such individuals. There's a lot to say about billionaires funding these projects, but I'll leave that to someone else.
One answer from a university research perspective is that the system motivates grad students to work on easier problems so they can get their Ph.D. in a reasonable amount of time. Spending 30 years solving a problem is not an option.
More established researchers don't do research, they look for funding to fund their grad students. Some biomedical labs do have specific diseases as research targets, but research in a direction often stops when the grad student graduates and new grad students tend to favor new problems, not continuing work on someone elses problem. Research funding is controlled by the older, more established researchers and they mostly go with what they know, rather than trying wild new things, even if little progress is being made.
In the social sciences things like hunger, health, shelter are somewhat studied but most researchers find more money in studies of how to motivate more workers to higher productivity and similar corporate aims. Mostly it's only governments who fund studies on the important problems because there is no obvious money to be made from the results, even though the societal benefits would be huge. War eats a lot of government funding.
There are a lot of other reasons also, like the financial industry grabbing programmers and physicists who could be working on more important things instead of gaming markets. I suggest taking a look at the buildings in your town, the nicest, biggest, most expensive buildings are where the money is going. The oldest, most run down places are where it is not going.
More established researchers don't do research, they look for funding to fund their grad students. Some biomedical labs do have specific diseases as research targets, but research in a direction often stops when the grad student graduates and new grad students tend to favor new problems, not continuing work on someone elses problem. Research funding is controlled by the older, more established researchers and they mostly go with what they know, rather than trying wild new things, even if little progress is being made.
In the social sciences things like hunger, health, shelter are somewhat studied but most researchers find more money in studies of how to motivate more workers to higher productivity and similar corporate aims. Mostly it's only governments who fund studies on the important problems because there is no obvious money to be made from the results, even though the societal benefits would be huge. War eats a lot of government funding.
There are a lot of other reasons also, like the financial industry grabbing programmers and physicists who could be working on more important things instead of gaming markets. I suggest taking a look at the buildings in your town, the nicest, biggest, most expensive buildings are where the money is going. The oldest, most run down places are where it is not going.
Because people have bills to pay, and income is not correlated with impact on society.
Because they are hard?
Hard to achieve, every pioneer had to shed blood sweat and tears, little or no profits in the short time (1-5y), and therefore the ratio of effort vs remuneration is not as favorable as joining a well-established cash-cow.
If you don't mind me asking, what makes you ask?