š® Information in war; ChatGPT and twitter; old scientists; life on Mars and radical millenials++ #404
An exponential tale: Ukraineās supremacy in information makes its military more effective than Russiaās reliance on explosives
Hi, Iām Azeem Azhar. I convene Exponential View to help us understand how our societies and political economy are changing under the force of rapidly accelerating technologies.
In todayās edition:
Ukraineās supremacy in information helps its military compensate for Russiaās advantages in things that go boom
Using ChatGPT to explore counterfactual scenarios,
The age of innovators changes how creative they are, and whether their research is more theoretical, or experimental.
Sunday chart: Information & energy
We are witnessing exponential growth in the number of active satellites, even excluding the 3,335 Starlink satellites. It is not inconceivable that there will be 50,000 satellites circling the globe in a few years ā with all the attendant hazards of space debris and widespread cyberattacks.
The Economist analyses the impact of the Starlink satellite communications network in Russiaās invasion of Ukraine. āUkrainian soldiers upload images of potential targets via a mobile network enabled by Starlink. These are sent to an encrypted group chat full of artillery-battery commanders. Those commanders then decide whether to shell the target and, if so, from where. It is much quicker than the means used to coordinate fire used up until now.āĀ
This real-time communication channel shows how information can supplant energy in a given process (in this case, meting out hell to violent invaders). Russia is outpacing Ukraine in delivering explosives, but Ukraine, augmented by the precision afforded by information, is, in many ways, being militarily more effective.
You can trade brute-forcing with energy for the precision that information and computation affords.
This substitution, information for energy, is common amongst processes of the Exponential Age. Precision fermentation applies information, amongst others, to biological processes (like making consumable protein for humans) and does so more efficiently than creating processes in large mammals. In silico, we can simulate a thousand trucks driving a billion routes to find better routings. Information replaces the heavy energy cost of actually driving those routes.
Of course, information still needs physical stuff (microorganisms, artillery tubes, bioreactors, high explosives, people) to actuate in the real world. We are not fully dematerialised. Yet.
Key reads
ChatGPTās take on Twitter v. Mastodon. EV member Nicolas Granatino interviewed OpenAIās ChatGPT about how decentralised social media platform Mastodon compares to Twitter. Iām amazed at how well it played out some counterfactual, āWhat if?ā scenarios. (We did some similar experimentation). E.g. what would happen if the Twitter graph was opened to organisations to build, and the filtering algorithm was offered to users? (OpenAI might sell existing shares in a tender offer that would value the company at around $29 billion, making it one of the most valuable American startups).Ā
The age(s) for innovation. Scientists and inventors tend to become less creative with age, as they become increasingly specialised and rely on older concepts. Younger professionals rely more on frontier knowledge. Interestingly, āit seems that Nobel prize-winning conceptual/theoretical breakthroughs really are clustered among the young, and experimental ones among the old.ā See also, papers and patents are becoming less disruptive as time passes (this is a fascinating and important read). This suggests researchers need to focus less on the quantity of publications, and spend more time learning about new and cross-disciplinary concepts. Is this a function of broken research funding models? Could new AI tools help researchers jump across disciplines more easily?
A better way to understand climate scepticism. Automated classification of content into either ātruthfulā or āmisinformationā can be misleading. Research looking at the structure (the networks of interaction between users) instead of content shows that discourse in climate-contrarian echo chambers is more focused on climate action delay than the outright climate denial.
Azeemās commentary: Climate scaling
In tomorrowās commentary, I will explore the question I still puzzle over: how do we rapidly scale the climate tech we need for climate change mitigation and transition?
Market data
By 2030, oil and gas supermajors still plan to allocate 84% of their capex to business-as-usual (aka fossil fuel) projects, according to New Financial.
Norway does it again: 79% of cars sold in Norway in 2022 were electric.Ā
The median B2B SaaS startup takes 10 years to exit. The median consumer Internet startup only takes 7.
Meta and Googleās dominance of US advertising has fallen to 44.9% of US market share (from its 2017 peak of 54.7%).Ā
VR headset shipments dropped 12% in 2022.
The hoarders down under: Australians are spending $110 per month to store stuff they donāt need.
Short morsels to appear smart at dinner parties
šA long read on why sending humans to Mars would be impractical and wasteful, and also rendered useless by advances in technology, via EV member Paola Bonomo. Follow-up with Asimovās quasi-philosophical short story āLittle Lost Robotā.
š Johns Hopkinsā interactive map of the universe.
šøļø Webs as an extension of spidersā minds.
𩺠Google and Deepmind released MedPaLM, a large language model designed for the medical field.
šØ Both American and British millennials are not following the age-old trend of becoming more conservative with age. (Via Morten N. StĆøstad)
Exponential knowledge community
šļø We are hosting a private briefing for members on trends that will shape 2023. Join me on January 26, 9-10 am GMT or 4-5 pm GMT. *The event is open only to members of Exponential View with an annual subscription*. You can RSVP below:
š We asked members to share their 2023 predictions with us. Hereās a selection of 18 wondrous forecasts from our community ranging from biotech, AI, markets, to climate change mitigation, medicine and global politics.Ā
š¬ What members are talking about in our Slack space: NFT hype stretched out thin, citizen assemblies in Europe tackling biodiversity loss and end of life, reading to take you from 0 to 1 to understand deep tech. To join us on Slack, fill out this form.Ā
See also:Ā
Heather Myers published a piece in the Harvard Business Review about the āreal-worldā testing she does to support innovation.
Vess Ignatova writes about the role Biz Ops teams play in developing company strategy.Ā
Gianni Giacomelli covers the pandemicene, and human-movement physics in the January edition of his Collective Intelligence digest.
Share what youāre up to here.Ā
End note
Iāll be speaking at DLD in Munich from Jan 12 to 14th. Ping me if you will be attending.Ā
Iāll also be at Davos presenting my horizon scan for the coming years. It covers AI, electrification, deep tech, deglobalisation and strategic competition, catalytic government and the shape of democracy.Ā
Iām also talking inter alia on generative AI and a couple of other things. Ping me on the eventās internal messenger if youāre attending.
Cheers,
Azeem
Regarding the point "Age(s) of innovation": not sure if I red the paper after seeing it mentioned on EV or in some other source. But the chapter 3.1. of the paper "TALENT VERSUS LUCK: THE ROLE OF RANDOMNESS IN SUCCESS AND FAILURE" ( https://www.worldscientific.com/doi/10.1142/S0219525918500145 ) shows very interesting suggestions (and simulation results) how funding for scientific research can be improved. And, at least from my perspective, the correlation between "randomness", "serendipity" and "theoretical / conceptual breakthru" makes intuitively sense.
The gist in terms of funding approach: give chance a chance - instead of trying to reproduce hindsight success.