💡E03: Three Overrated Technologies from the Future Fusion, Bitcoin & LEO Internet + Ultrasound for Hibernation; Wind Power; Skeletal Editing; Carbon Pricing; & Psychedelics++

Spotting a hype cycle: Why nuclear fusion, bitcoin, and LEO satellite Internet are overrated. Don't @ me.

Jun 08, 2023

It’s easy to find reasons why things will fail. You can’t be an entrepreneur or investor and not be optimistic. Things that don’t exist today, will exist tomorrow. It’s that optimism that makes the startup ecosystem exciting, and why lots of young people want to do it. You can say you are building the future. Or hodling it, if that’s your thing. And that’s great.

But not everything succeeds. And more importantly not everything can succeed *now*. And that’s important. The *now* is the hard bit. Pets.com and Chewy.com. Webvan and Instacart. If you would have said in 1999, yeah like online delivery is inevitable, but it’s probably not viable with today’s technology and customer comfort with e-commerce. You might have been called a skeptic. Sure, but you were right. Man the things I said about 3D printing in 2012…

It’s easy to say X is the future, everyone thanks you for your optimism. I mean whole brain emulation is the future. On some timeline. I can say deep geothermal energy, fertility technologies, and brain recording technologies are undervalued and people say, fair beans. But when I say X is overrated, well that’s more problematic. As an investor, I could just not invest in that I suppose. No need to shout about it. But as a researcher, I dunno, is there some/any value in saying: other people are doing it wrong?

So nuclear fusion, LEO satellite Internet, and Bitcoin, then. These are interesting cases and I have slightly different analysis. Fusion is still too early for venture capital. LEO satellite will remain niche for a long time. And bitcoin. Well, it’s a *successful* revolution and meme, but, unlikely to be successful on its own terms. Would you rather have started a revolution or be rich?

✍️ Three Over-valued Technologies from The Future

Fusion: $3.3 billion in venture capital can’t be wrong, can it? Can it?

Despite perennially being 30 years away, 2022 was a breakthrough year for fusion. The Joint European Torus (JET) project sustained a five-second reaction generating 50 million joules. And The Lawrence Livermore National Laboratory (LLNL) achieved ignition (a self-sustaining reaction). This followed a breakthrough in 2021 in which MIT and Commonwealth Fusion Systems (MIT-CFS) broke the magnetic field strength record with a new high-temperature superconductor. There has been talk of commercial reactors in the 2030s from startups after they raised a record $3.3 billion in 2021 and 2022. Clean, abundant energy is on the horizon, right?

Wrong. Like honestly, I am so glad startups are working on fusion. And maybe, people have to say things to get the money they need to make those things actually happen. I know how things work. But still. It will be all the years before we have an operational plant. One of the biggest challenges is that we don’t even know how to build a reactor. We don’t even know which materials to use. During a reaction, neutrons are released, damaging the reactor's walls. Known as the "first wall problem,” solving it requires new and expensive materials. Known materials, such as tungsten or beryllium, are incredibly expensive. And that’s just the walls. We also have no idea how to build the components that will run the reactor, like the plasma injectors and radio frequency heating systems. And that’s before the fuel problem. Fusion reactors will use tritium produced from lithium-6, an isotope that is only produced in large enough quantities in North Korea, China, and Russia — for use in nuclear weapons. Building facilities for testing new materials and components would take the best part of a decade if we started today. And we aren’t.

The bigger question is just how useful nuclear fusion will be in the 2040s anyway, with another two decades of cost declines in solar, wind, and battery storage combined with what looks like a faster pace of small modular reactors for baseload energy, and I hope deep geothermal energy. But ignoring all that, let’s boil it down. Venture capital firms have 10 years (maybe 12) to exit their positions. Ignoring all the detail, you have to ask, how plausible is it that a startup can build an operational reactor by 2033? As noted last week, the consensus is that fusion is 17.8 years away, so 2039/2040. Maybe you can bend the curve a little, but by more than 5 years? Maybe in software, but in this case we have to dig stuff and build it and then get regulated. Colour me sceptical.

More [[Nuclear Fusion]]

LEO Satellite: Internet for the 37 percent? Nope.

The potential of the Low Earth Orbit (LEO) satellite Internet is hyped beyond reality. LEO satellite Internet is delivered by satellites orbiting up to 2,000 km above the earth’s surface, compared to traditional satellite communication from geosynchronous earth orbit (GEO) satellites at 36,000 km. GEO is where you get your GPS, DIRECTTV (for our American friends), or Sky, for my European partners. LEO being closer to the Earth, achieves much faster throughput (approximately 100Mbps) and much lower latencies (less than 30ms). LEO has the potential to speed up Internet speeds for those currently underserved. A good datapoint is that the average download speed globally is just 30Mps, and 37 percent of the world's population – or 2.9 billion people – have never used the Internet. 2.9 billion people. Just think about that for one moment. *We* are all out here talking about ski googles and AGI. And nearly 3 billion people have never used the Internet. Your TAM is smaller than you think. Anyway, LEO satellite is amazing for people using GEO satellite internet today, war zones, and people and machines in remote areas.

But, predictions that LEO will disrupt the telecommunications industry by beaming the internet to households and cities worldwide just doesn’t stand up to reality. A high-impact case must argue that LEO satellite internet will deliver Internet more cost-effectively than 5G and 6G cellular and wireline fiber. The UN ITU found 95 percent of people in the world could theoretically access a 3G or 4G mobile broadband network, but affordability is the primary barrier. Onboarding the last third of the population is a development challenge, not a technical challenge. If cost is the main driver to adoption, even if Starship can get close to launching for $100 per kilogram, I struggle to see how beaming Internet from space can compete with 5G, especially mmWave 5G, WiFi 7, and even 6G in the early 2030s. I predict that by 2030 LEO satellite subscribers will be roughly the same as in 2022 (2.5 million).

I did see the Hubble Network announcement, a bluetooth network for IoT announcement. Thinking about this has been clarifying. I realise I am less bearish on a LEO network for machines and objects than people. Even with a full constellation the Hubble are aiming for 15 minute latency, so they aren’t competing with 4G, 5G, WiFi, etc. For applications where latency and bandwidth can be on the low side, a “cheaper” network makes total sense. If of course, LEO bluetooth networks are indeed cheaper.

More [[LEO Satellite Internet]]

Bitcoin: A successful revolution but a failed coin(?) (i’ll add a ? to hedge my bets)

Yeah sorry team. I don’t relish writing this. I have previous as they say in the UK. Bitcoin has been incredibly successful. Maybe not against its original goal of becoming a peer-to-peer payment network. But certainly in terms of a robust store-of-value network. It could also be considered successful purely as a practical example of a blockchain, ushering in a new age of cryptographic and distributed systems exploration. From this, we almost certainly awouldn’t have progress in things like zero-knowledge proofs, distributed file storage, and distributed computing networks. These tools will be a core part of the next generation of Internet technologies. It will also go down in history as the catalyst for the faster adoption of new technologies in the broader banking system, making the traditional system faster, safer, and private (depending on central-bank digital currency design).

Revolutions aren’t necessarily failures if they fail to overthrow the power structures immediately. The revolutions of 1848 in France, Hungary, and Ireland were largely unsuccessful but sowed the seeds of nationalism and liberalism. When seen through this lens, Bitcoin sowed the seeds of decentralization and individual rights in the digital realm. Bitcoin as an idea will persist, even as Bitcoin as a technology largely stalls. To be successful, Bitcoin must prioritize censorship resistance above all else, including usability and programmability. What makes it compelling as a non-political store of value also makes it difficult to improve. Alternative networks offer more effective privacy (Zcash), more programmability and functionality (Ethereum), and faster transaction settlement (Solana, Polygon). Central-bank digital currencies and digital wallets likely win over mainstream consumers who don’t care about privacy or self-custody, prioritizing convenience and features. There will be an important niche as a hedge against censorship for both goodies and baddies. It needs to exist, but I see very few pathways for Bitcoin to achieve mainstream adoption or even get back above $60k per coin ever again. Bitcoin is a successful meme but an unsuccessful technology.

Why might I be wrong? Any bull case probably relies on the success of the lightning network and bitcoin becoming a cheap payment network competing on cost and convenience versus existing payment networks. But fundamentally, any case for Bitcoin is a case against the Government. I mean the technology is rooted in a libertarian value system. I think Bitcoin is a political-technical solution to a political problem. The banks caused a financial crisis. Everyone was mad. No-one went to jail. Then you have all these unbanked because the financial system has failed them. And then there are greedy monopolies. Web2 surveillance capitalism and all that. Sure. But you know what else can solve these problems? Effective Government. I appreciate some people will thin that’s a contradiction in terms. But as I said up top, I’m an optimist. And my optimism extends to politics too. I think many engineers look at politics and think it’s ineffecient, slow and corrupt. But by and large that’s because the fair allocation of resources is a hard problem. It’s an NP-hard problem. I think we have a pretty good financial system that works for *most* people. If you are in Argentina, maybe not. But that’s a political problem not a technology problem. Can we make it faster? Absolutely. Can we make it more private? Yes. And can we make it cheaper so more people can be banked. Yes. And all of those things can be done without Bitcoin.

More [[Bitcoin]]

❌ Mailbag

We had some good fun this week over on Linkedin.

On fertility technologies: The Economist wrote a leader on the problem of declining fertility. To some extent validating the point that fertility technologies are underrated.

@Arnaud de la Tour of Hello Tomorrow: “Isn't that a good thing that fertility declines, given that we already need more than one planet to sustain our society?”

@me: “Quite a difficult discussion and maybe not the best format on LinkedIn but…I have a feeling where you come down on this debate depends on the relative weight of the negative impact of climate change versus the negative impact of slower/no economic growth. Both have devastating consequences, but I think we come down slightly on one side versus the other.”

Assumptions: 1/ More people = worse climate change (this is generally true, but decarbonising the economy/lower animal consumption would reduce impact) 2/ Fewer people = lower economic growth (this is generally true, but AI might replace some humans as labour untethering this relationship)

I think we should focus on addressing climate change through decarbonising the economy not by reducing the population. Lowering economic growth would reduce our ability to address climate change anyway. I also don't see the political pathway for anti-natalist policies. Almost all countries outside of sub-saharan Africa have issues in managing an increasing aging population with fewer workers. If anything, pro-natalist policies are increasing as with China and Hungry.

@Saurabk Tak: “first fertility is a very direct measure of health. Decreasing fertility means increasing health issues most probably because of the lifestyle diseases. A side point here that natality should not be confused with fertility. We control or can control natality, fertility controls us.”

This point of declining fertility as a lifestyle disease was a great point. Reminds me of the Chemical Hunger series from SLIME MOLD TIME MOLD.

I had quite a few conversations with some other people in DM, and I learned any discussion of fertility is strongly tied up with the climate crisis. And that we need to correctly weight the likely pushback on any pro-natalist policies. I find it hard to imagine a political party winning much support at the moment with a: we need more people! Even if it’s true.

And over on Substack Chat:

on geothermal:

@me: “ Yeah right, I’ve asked everyone I know in the energy industry, what am I missing here? Am I totally underestimating the difficulty of using mud motors, rotary steerable systems, and underreamers? Or the political challenges of “fracking”? It seems like all it would take would be a strategic decision by an oil major

@Neil Swaelens: “The challenges primarily relate to the drilling mechanism. Cost go up exponentially when you are dealing with hard rock. Non-friction drilling is hence the way to go. The other challenge relates to the conditions in the borehole at great depths (i.e. pressure/temp). So the less complex the system in the borehole the lower the chance of issues arising. I think the conditions surrounding fracking are well understood yet the political angle may indeed be a hurdle --> education is def required in those circles. I think the other challenge to it from a VC angle is the general reluctance among climate VC to perhaps work together with oil/gas due to perception issues yet I may be wrong here. Some amazing work is being done by Nabors in the space. Also great companies like GA Drilling and Quaise.”

fwiw, Quaise raised a monster $52M Series A last year. Saying: “There is no energy transition without deep geothermal. Deep geothermal offers an always-on, inexhaustible energy source without fuels, without waste and without geographical limitations. It is also the most powerful, reliable and abundant primary clean-energy source on Earth.”

🗞️ News

🗞️ Ultrasound Puts Animals into a Curious Hibernation-Like State

“If we can identify ultrasound-sensitive [proteins] in other parts of the brain, we may modulate other behaviors,” she says, adding that what those might be remains to be seen.” » I’m closely tracking light [[optogenetics]] and sound as non-invasive modulators of cells. Strong intersections here with BCI and obviously human space travel. Imagine EarPods or the Vision Pro modulating behaviours directly through the ear or eyes via ultrasound...

🗞️ ‘Almost magical’: chemists can now move single atoms in and out of a molecule’s core

“An explosion of skeletal editing methods to insert, delete or swap individual atoms in molecular backbones could accelerate drug discovery.” » I didn’t have skeletal editing as a technology in the tracker. Added and will complete soon. Also very cool name. Better brand name than CRISPR or prime editing imo.

🗞️ Psychedelics promote plasticity by directly binding to BDNF receptor TrkB

“Our data confirm TrkB as a common primary target for antidepressants and suggest that high-affinity TrkB positive allosteric modulators lacking 5-HT_2A activity may retain the antidepressant potential of psychedelics without hallucinogenic effects.” » I wrote “A high-impact scenario sees the neuroplasticity and neurogenesis thesis proved, which would broaden the clinical application.” Also, this work hints at a future where psychedelics can be taken in pill form rather than as a treatment in a therapeutic setting. For those reasons, this is a very important finding. [[psychedelics]]

🗞️ Global wind turbine orders surge to record in first quarter, led by China

“Global wind turbine order intake surged by 27 per cent in the first three months of 2023, setting a record for the first quarter.” » Just keep revising your renewable deployment forecasts upwards I suppose. Pair this with the news that following two consecutive years of decline, onshore wind capacity additions are on course to rebound by 70% in 2023 to 107 GW, an all-time record amount. And on the solar front: “Global renewable capacity additions could reach 550 GW in 2024 in our accelerated case, almost 20% higher than in the main forecast.”

Bad news for SMR deployment though, the original target price from NuScale was $55/megawatt-hour (MWh) and is now $89/MWh. This includes hefty subsidies to the tune of $4 billion from the Department of Energy and the estimated $30/MWh subsidy in the Inflation Reduction Act (IRA). Hard to bet against wind, solar and batteries in the race to decarbonise. [[Wind Power]]

🗞️ State and Trends of Carbon Pricing 2023

“The share of global emissions covered by carbon taxes and emissions trading systems (ETSs) has grown from 7% to around 23%” » Important number to track. Generally speaking, the broader the coverage the faster renewable adoption. Worth tracking the price too, obviously that drives adoption within the ETSs. Pretty much all of the cleantech assessments note ETS coverage and carbon price as key drivers of adoption. [[Solar]]

☎️ Interviews on Privacy-Enhancing Technologies

A couple of weeks ago, I gave a talk for the PETs Demo Day conference organised by the UK Centre for Data Ethics and Innovation and the US National Science Foundation (NSF). I had a framing breakthrough:

TLDR: We are in a difficult Phase 1 for PET adoption. Adoption speeds up in Phase 2 with the shift from consulting-led sales to product. Phase 3 is a cloud-level opportunity (I wrote a $1 trillion opportunity which I still stand behind)

I though it was time to re-up, the privacy-enhancing technologies and collaborative computing thesis and some of the follow-up interviews:

☎️ Christine Huang, Data Privacy & Protection at SAP #001: On why data-sharing is cultural; global data-sharing rules are unlikely; and a future tech stack won't be edge vs cloud

☎️ Flavio Bergamaschi, Private AI and Analytics at Intel #002: On why data integrity is just as important as confidentiality; why the importance of crypto-agility is overlooked, and the five groups you need to convince to sell data collaboration software

☎️ Dr Hyoduk Shin, Professor of Innovation at UC San Diego #003: On realigning incentives towards data sharing; why culture will be the most important driver; and why we will end up with a globally fragmented data economy

☎️ Stijn Christiaens, CEO of Collibra #004: On competing against non-consumption; why Data Asset Systems of Records are like CRM systems 20 years ago and; the changing role of the Chief Data Officer

☎️ Rick Hao, Partner, Deep Tech at SpeedInvest #005: On more data as the driver of PET adoption; why healthcare needs its own infrastructure; and why machine learning is likely to be a catalyst for adoption

☎️ Jordan Brandt, CEO of Inpher #006: On why monolithic incumbents are going to be left with all the data risk; why taking a vertical go-to-market strategy is sub-optimal; and why data federation is an inevitability

State of the Future

Discussion about this post