A Technologist's Guide to the Future | Nathan Myhrvold | Knowledge Project Podcast

Unlike pandemic or military emergencies where rapid response can immediately stop new harm (new infections, battlefield casualties), climate change has a multi-century lag where stopping emissions today only stops future warming increases, but past warming persists—making climate uniquely intractable to emergency response because action today cannot remedy past emission consequences.

“so suppose that we Panic one day and we say right from Tomorrow onward there will be no more emissions well unfortunately the chemistry and physics of uh climate change is not very forgiving so when you omit CO2 it stays in the atmosphere for a really long time um thousands of years tens of thousands of years”

The Air Force historically resisted drone adoption despite drones being superior to human pilots because the Air Force is staffed by pilots and former pilots who have institutional and identity-based resistance to being replaced by machines, showing how institutional self-interest can block superior technology adoption.

“it's been true for a long time uh that um humans are the weakest part of a fighter jet yeah you should use drones yeah you should totally use drones but the Air Force which is composed entirely of Pilots or former Pilots hated that idea um but drums are here to stay and there will be um aerial combat drones”

Autonomous systems (autopilots, auto-landing systems, modern trains at 300 km/h, subway systems) are significantly safer than human control because humans cannot operate fast enough, are unreliable, and are easily distracted, yet human crews remain on these systems for historical/safety backup reasons rather than performance reasons.

“there's no modern jet that doesn't have an autopilot for God's sakes um so uh and of course that's because in that particular application um computers do a way better job”

Ecological/environmental problems are rapidly solved when they are localized in space, localized in time, and responsibility is easy to trace (Love Canal, acid rain, smog in Los Angeles)—but climate change represents the worst case scenario because it is diffuse globally, occurs over decades to centuries, and causation is distributed across all economic actors, making rapid collective action nearly impossible.

“whenever we have a problem that is an ecological problem that is localized in space localized in time and is easy to trace responsibility we fix it Love Canal United States there's a Superfund Site huge amounts of chemicals were dumped for many years by a company made the ground horribly poisoned it's a tragedy it should never have happened but it did happen so okay we fixed it um we even fixed a slightly less localized issue with you know most Aaron pollution and most um water pollution United States is solved”

Global cooperation on climate and energy is difficult because developed nations (US, UK) became wealthy by burning fossil fuels for 150+ years, yet demand developing nations (China, India) restrict their own emissions despite lower current living standards—creating a moral hazard and incentive problem where developing nations face demands not to follow the same path that enriched the West.

“you have this issue that the developed countries the world the United States the UK is an example we've both been burning lots of fossil fuels for 150 years at least um and yet we want to tell China and India what to do that they should not do that yeah and those people say hey what do you mean we have we have uh people who are living way below your standard of living yeah you got rich this way it's our turn damn it and irrespective of what you think about the moral imperatives there short-term trade-off versus long term and so forth we just don't have a way to force them yeah unless you want to go to war over it and that would be insane”

Bioterrorism is potentially more dangerous than nuclear weapons because nuclear weapons are difficult to develop (few countries have them) while bioterrorism is cheaper, easier, and harder to control—making biological weapons a more likely asymmetric threat from actors with ill intent.

“but it can do cyber you can do biological weapons you can do other things yes um this will sound very strange but thank God for nuclear weapons in that most um countries with ill intent really they they see the United States or uh the Soviet former Soviet Union now Russia as superpowers nuclear weapons they want a nuclear weapon and it's really hard to make a real nuclear weapon and it's so hard that even though a few countries have gotten them it's not like everybody has them and it's cheaper it's easy uh bioterrorism is potentially worse than that because it's potentially much cheaper and much easier and uh it's also much harder to control um that whole thing makes it a way worse thing so I think we'll see a bioterrorism event someday”

Historical precedent shows that when humans make major technological transitions to solve one problem, we create new problems—agriculture solved food scarcity but created famine catastrophes; automobiles solved horse waste/urban pollution but created traffic deaths and air pollution—suggesting geoengineering will have consequences but they are preferable to unmitigated climate catastrophe.

“yes exactly just like all of human history yeah um you know it we were hunter-gatherers then we invented agriculture which was great we could have a larger population the population didn't have to migrate with wherever the food was so we could have a civilization we had Leisure Time We Invented art and culture and all these wonderful things flourished under agriculture but then we discovered oops there's things called droughts and crop failures and and so for the first time you had catastrophes for hundreds of thousands or maybe even millions of people died you can't have a million people die at once and under gather a thing you don't have enough a there's not enough people and B they're all dispersed and they all have their own situation there then you could so yeah agriculture was terrible but then we've learned to fix those things in agriculture we've learned to fix more and more and Away you go”

Climate physics creates an irreversible lag: even if humans stopped all CO2 emissions immediately on day one, the climate would continue warming for approximately 85-110 years, then take another 145 years to return to baseline temperatures—meaning that by the time society panics about climate change and reduces emissions, centuries of additional warming are already locked in by past emissions.

“so when you omit CO2 it stays in the atmosphere for a really long time um thousands of years tens of thousands of years and people argue but you wouldn't be totally wrong by saying there was a twenty thousand year life span now that's the same as never from a human life I'm perspective so turns out if you stopped all emissions on day one the climate will continue to drip drift upwards for about a hundred years I've written a bunch of papers in global warming I've done this calculation other people have done this calculation you know you can equivalist whether it's 85 years or 110 years but it's something in that range then it'll slowly start going down it will take 145 years for it to equal the same temperature on the day that you panicked and shut it off”

More people are killed annually by car accidents in the US than by 9/11, and roughly double the number of people murdered, yet car deaths receive far less outrage or policy attention because society has adapted psychologically to car risk in exchange for the freedom and mobility they provide—suggesting humans make complex risk-benefit tradeoffs that are not always rational.

“there's more people will die in uh car accidents the United States this year then uh died you know by like a factor of 10 died in 9 11. um usually the number of people killed or maimed in car accidents is about double the number of murders we're outraged about murders we're not that outrage about car accidents yeah and partially that's because we're addicted to the freedom of Mobility that cars and other Motor Vehicles give us”

Fire sprinkler systems represent rational risk management for low-probability events: each building has a low individual probability of burning, but millions of buildings exist, so statistically buildings burn daily—therefore installing sprinklers everywhere is economically justified, providing an analogy for pandemic preparedness policy.

“my example there is um fire sprinklers so if we look up in this building or nearly any building that you would go in that's a modern building and most older ones even you see fire sprinklers yeah but you also see lots of 100 year old buildings if you're in Seattle on the East Coast you see 200 year old buildings so the probability of a burning of a building burning down is not very high but we have millions of buildings so every day some building burns down and uh fortunately it's not a coordinated thing it's just one building will burn down or maybe you have a tragic wildfire that like destroys a whole town has occurred in the mountains in California a few years ago those fires happen often enough that even though you have a low risk on any individual day we say right this is it's it could happen we're gonna put sprinklers everywhere”

Humans consistently prioritize diseases that kill people immediately (cancer, heart disease, HIV) over low-probability high-consequence events (once-per-100-years pandemics), showing how resource allocation reflects availability bias and recency bias rather than expected value calculation.

“I had some very reasonable people who would argue with me on this pandemic issue they would say but look I study heart disease or cancer or HIV or some other terrible disease and yeah it may take you know we may have a pandemic at some point but shouldn't we put our first priority towards diseases that kill people every day rather than disease or situations that occur once every 100 years”

Early electric vehicles charged on the Eastern US electrical grid actually increase climate harm rather than reduce it because the grid is primarily powered by fossil fuels, though Washington State EVs have lower climate impact due to nuclear and hydroelectric generation—demonstrating that EV climate benefit is location-dependent on electrical grid composition.

“if you drive a Tesla and you live in the East Coast United States statistically speaking you're hurting the climate not helping it Washington State we can feel slightly more slightly better about it because we have a mix of nuclear and hydro so most of the power in Washington state is at least clean so if you charge your Tesla up there that helps”

Decision-making under uncertainty requires two components: (1) being analytical and careful to reach good initial decisions, and (2) monitoring outcomes and being willing to change course when evidence indicates the decision was wrong—this framework applies to both personal and organizational decisions.

“I think there's there's two things one is you try to be as analytical and as careful as you can about making a decision but the second is you also have to monitor how it's going and not be afraid to change your mind”

Humans are neurologically wired to prioritize immediate gratification over long-term consequences, which is why people fail at dieting and exercise despite knowing the long-term benefits—this cognitive bias makes addressing long-term civilizational threats like climate change extremely difficult because the payoffs are distant and diffuse while the costs are immediate and concentrated.

“humans are not great at long-term things Okay so um we all know that we should eat right and work out lots of us don't and that's because the immediate gratification of oh I'm late for an interview this morning I I gotta leave I I don't have time to work out that was me this morning or wow that donut looks really great or there's an extra slice of pizza left man we are such suckers for that it's ridiculous”

Even if new renewable energy capacity grows rapidly, the percentage of total global energy from renewables remains constant or shrinks because global energy demand is growing faster than renewable capacity additions, especially as China and India build new coal plants—so measured CO2 concentrations continue rising despite renewable growth.

“the trouble is the world's also growing economically which is a good thing mostly but as a result the percentage of Renewables even if we grow Renewables really fast it's hard to grow fast enough that it makes up for all the coal plants that are being built in China and India and other places whose economy is booming um and if you measure CO2 up on the top of Mauna Kea and Hawaii that's where the uh the best measurements come from every year it's higher so no we made no progress”

Microsoft's founding vision of 'a computer on every desk and in every home' was criticized as an absurd fantasy in the 1980s-90s, but has been vindicated by smartphones and computing embedded in vehicles and infrastructure, establishing that long-term technological forecasts can be accurate despite initial ridicule.

“you know what Microsoft we used to have a slogan for the company a computer on every desk and in every home and I took so much grief from people when I would say this in speeches where they would say oh I'm never going to have a computer in my home that's just an absurd nerd fantasy well it worked now you have one in your pocket”

World War II demonstrated that the United States, despite being isolationist and unprepared before Pearl Harbor, rapidly mobilized enormous military and industrial capacity, with half of Army officers replaced within 3 months due to being 'parade generals' (unprepared for actual conflict)—showing that emergency mobilization can rapidly replace institutional inertia.

“in World War II the America had after World War one was tremendously isolationist we neglected our defenses we but then there's Pearl Harbor and that was a wake-up call that you just couldn't ignore most people don't realize we had not declared war on Nazi Germany uh before Pearl Harbor and in fact we never declared war on them first within a few days of Pearl Harbor Germany declared war on us sounds like oh right we have a multi-front war and America got serious um half of the officers in the U.S army were let go within three months of that because they were what people call parade generals they were a bunch of maybe distinguished people but all they had done for the last you know 30 years is lead parades and marching bands um”

Natural gas is worse for climate than coal when accounting for methane leakage in extraction and transport, because methane has 120x the global warming impact of CO2 in the first year after release, and does not require very much leakage to make natural gas worse than coal overall—contradicting the gas industry's claim that burning natural gas produces half the emissions of coal.

“Natural Gas actually produces more global warming in in almost every case than coal um the the gas industry loves to say well when you burn it it's half of coal um and that's true but of course you also spill a lot of methane into the air methane on the first year you spill it it's 120 Times Higher a global warming impact than CO2 over time the methane decays into CO2 where it only becomes just as bad as CO2 um it doesn't take very much of a leakage of methane to make it worse than coal and all current scientific estimates say that net net natural gas is worse”

Geoengineering deployment should only occur when climate change becomes an 'unmitigated disaster,' analogous to how emergency open-heart surgery is only performed in life-threatening situations, not for minor ailments—establishing a proportionality principle for when technological intervention is justified.

“so you should only pull out the uh the geoengineering fix if it's otherwise unmitigated disaster just like they wouldn't crack my chest open if I went into the emergency room with a hangnail okay they Reserve that for a very special set of circumstances”

The manned space program was justified when human presence was the only way to explore certain classes of space missions, but this necessity ended long ago and today manned space flight has become primarily a reality TV show and space tourism venture (Blue Origin, SpaceX) with no compelling rationale for risking fragile human lives when robotic missions are far cheaper and more capable.

“the manned space program is another sort of absurd thing um it started off wonderfully that the only way we could explore um uh space What for a certain class of things was with humans um well it's long since past I mean today the manned space flight is a reality TV show um and in fact the The Cutting Edge of man and space flight is space tourism that's what um blue origin and SpaceX and others are gearing up for”

Humans systematically delay action on long-term threats until immediate crisis forces panic, at which point they mobilize rapidly—the pandemic response shows this pattern: warnings were ignored for years, but once deaths occurred in Seattle, society mobilized immediately to develop vaccines in 8 months rather than 10-12 years through conventional processes.

“humans will typically put stuff off and put stuff off and put stuff off and then you panic now we saw that with the pandemic okay I had been publicly saying that we had to worry about a both a natural pandemic and or a bioterrorism thing I've been saying that for 10 or 15 years other people have been saying it for longer than I had um you know in 1918 you have this horrible pandemic you know that killed you know three percent of the people five percent of the people worldwide something like that I mean that's a giant number it's a bigger number today than it was then and that was without airplanes and without us but no matter how much you pound the table and say we need to take pandemic seriously 1918 was a long time ago people forget”