The Current Thing: August 19
Also: how realistic is the fast unemployment AI scenario? Five friends of the show respond, exclusive for our Substack.
Happy Tuesday, we are live. The current thing is Chamath Palihapitiya’s new SPAC, the post of the day is Will Manidis’ take on “casino culture,” and our biggest gong is Databricks’ $100B valuation.
Here’s John’s run of show. 👇
Chamath is back with another SPAC
Chamath is back with a new SPAC focused on defense tech, AI, and DeFi. It’s U.S.-based but incorporated in the Cayman Islands. Sounds like a Silicon Valley mad lib — here’s what’s actually happening.
Chamath filed a prospectus with the SEC for a $250M special purpose acquisition vehicle. The money goes into a publicly traded shell company, which then acquires a “real” company and takes it public with less red tape and fewer guardrails than a traditional IPO.
Suffice to say, the timeline is not impressed. Chamath’s last SPAC wave (2019–2023) produced mixed results: 10 deals total, four liquidated (investors got money back, no target found). Of the six that closed, only SoFi trades above its $10/share launch (now $23). The rest cratered: Virgin Galactic is down ~99%, OpenDoor is down ~64% (but maybe about to see a comeback? Keith Rabois joined TBPN to talk about this last week), Clover Health is down ~75%, Akili is down ~96%, and ProKidney is down ~75%.
If you held an index of Chamath SPACs, you’d be disappointed. To his credit, Chamath isn’t telling anyone to YOLO in this time. He straight up says:
“We believe that retail investors should only participate if (a) this investment is a small part of an otherwise diversified portfolio, (b) this investment is a quantum of capital they can afford to completely lose and (c) if they do lose their entire capital, they will embody the adage from President Trump that there can be ‘no crying in the casino.’”
I believe this is the first time “no crying in the casino” has ever appeared in an SEC filing. So yeah, this is a game of poker, and you can choose to bet with or against Chamath. The bull case is that he probably learned a lot from his previous go around, he stayed in the arena, he kept a public profile, and there’s probably another SoFi-like company out there right now if you look hard enough.
The bear case is probably just that the IPO window is wide open, so there’s adverse selection among companies that can’t use the traditional IPO pathway. That being said, lots of people raise reasonable critiques of the current IPO process, so it’s not that crazy to have multiple options. Investors should just work to value the underlying business of the SPAC’d company like any other investment. Build a DCF people! Or maybe channel Charlie Munger (RIP) and buy a wonderful company at a fair price.
Today’s lineup
Astronomer Co-Founder and CEO Pete DeJoy at 1 PM
Pylon Co-Founder and CEO Marty Kausas at 1:30 PM
Eight Sleep Co-Founder and CEO Matteo Franceschetti at 1:35 PM
Basic Capital Co-Founder and CEO Abdul Al-Asaad at 1:45 PM
Medallion CEO Derek Lo at 1:50 PM
Convoke Founder Alex Telford at 1:55 PM
We asked tech heavyweights about the ‘fast unemployment AI scenario’
We’re trying out a thing where we ask our favorite posters in tech about ongoing debates in the discourse, publish their answers on our Substack, then hash out their answers on the show (sometimes with them on air).
For today’s question, we asked roon, Doug O’Laughlin, Andrew Côté, Aaron Slodov, and Bryan Johnson about what we’re calling the ‘fast unemployment AI scenario,’ based on this tweet by Nic Carter.
Here’s exactly what we asked:
Is AI going to take the value of labor to zero in a sudden shock scenario, or will its impact unfold gradually, with labor and traditional businesses proving more resilient than expected?
For context: VC Nic Carter recently argued we’re only a few years from a labor collapse and runaway capital loop — “Marx’s worst nightmare/fantasy.” BuccoCapital Bloke countered that the opposite is more likely: AI’s effects will take far longer, and labor/businesses will remain more durable than the hype suggests.
Responses below 👇
i actually wrote about this extensively, even before chatgpt. but my overall take is: there is a role for human labor into the far future due to comparative advantages.
i believe progress is mostly gradual, with certain shock events interspersed. there may be a moment soon at which models will actually encompass *all* rather than *some* of the core functions of an L3 software engineer, let’s say. and that will create a discontinuous labor shock. but it won’t be the end of labor. — roon
History often shows that the future comes a bit slower than first feared. Almost every new technology has had a bubble (of inflated expectations), then a slow actual impact.
While I don’t believe in a ‘fast unemployment’ scenario, I do believe that, on the margin, the issue is a real pressure. And remember: even subtle shifts in supply can produce enormous surpluses (e.g., it only took a 1 percent increase in shale supply in the 2010s to destroy oil price). So while I think we don’t all become unemployed quickly, things do matter heavily on the margin. — Doug O’Laughlin, President of SemiAnalysis
Technology takes jobs that suck and replaces them with jobs that are cool. This has always been true. What AI forces us to acknowledge is that today lots of jobs make you feel important, but are actually bullshit that waste everyone’s time.
Why is it impossible to build physical infrastructure? Because there is an army of 'environmental coordinators' whose job is to make everyone feel good about themselves, do 'stakeholder engagement', answer emails, and generate meetings. This is retarded and dumb. In the past almost everyone was a farmer, now almost no one is a farmer, yet there is more food than ever.
We have constructed a large society where the majority of people do not have to think for themselves to survive. Get a degree in accounting, be an accountant. Or law. Etc. With AI, any job that is script-following can be replaced. People will have to start thinking for themselves and making decisions. AI is bad at making decisions and good at recalling information. We will find that the net-net is, yes, many jobs go away, just like machines replaced physical labor.Now, everyone has to think for themselves. This is uncomfortable for a consumer economy predicated on hacking dopamine and wage slavery. "Labor" is another word for "person who is told what to do." It is a good thing for that to end, because it's not necessarily the case that most humans are unagentic idiots, just that we've bred them to be that way to work jobs that suck and dull the human spirit. — Andrew Cote, Founder of Hyperstition, Inc
I take the side of Bucco on this all day, every day.
AI is largely overblown by people in tech who don't spend enough time in the real world. We've created a very dull proto-form of AGI, one that with enough training data can perform a digital task.
I love the optimism of people in tech, but what they're so classically good at is creating something and assuming everyone in the world is going to drop what they're doing and use what they've built for the rest of time. And to some extent, this assumption has played out well for them: software is a commodity now, attention is an economy somehow (a poor reflection on our collective intellect). And yet, civilization still uses fax machines quite a lot.
Tech adoption looks dramatic because we usually look at the growth rate — not the total adoption itself — and even then, usually just the first layer of change (speed), without considering whether that speed is itself speeding up or slowing down. Rate of change of acceleration is more interesting to me. Most AI tools have insanely high churn because they capture you initially, then you put them down as they lose utility (aka they don't replace entire workflows for you).
The people who are firing workers are just replacing single-task labor, like customer service call centers. Like most of the history of civilization, AI is just another tool for humans to use, and the ones who learn to wield it first will dominate the ones that don't… until real AGI appears. — Aaron Slodov, Founder and CEO of Atomic Industries
Here is my take: No one knows what emerges with AI. We're accustomed to first principles thinking that allows us to make reasonable predictions about the future. That is no longer the case. We are now entering a new era that will require zeroth principle thinking where we humbly acknowledge we don't know. What we do know: no one wants to die right now. Don't Die is the next major, full-stack ideology that will help us practically function in this new world. — Bryan Johnson, Founder of Blueprint and Don’t Die
**Responses lightly edited.
In other news…
Eric Berger: China likely to beat the US back to the moon…
FT: Chinese podcast scene “booming”…
FT: Scientists develop brain implant capable of decoding inner speech…
Robinhood to launch football prediction markets in its app…
SoftBank’s Masa to invest $2B into Intel common stock…
MIT: “95% of organizations investing in generative AI have little or nothing to show for it”…
Cathie Wood’s ARK ETFs now taking in record inflows after years of underperformance…
Defense Sec. Hegseth introduces the “Pete and Bobby challenge”…
Thread: “Google’s reported search query growth may be AI-driven, not human”…
Google to build small modular reactor at Tennessee site…
NYSE relocates Chicago-based equities exchange to Texas, releases trailer…
Wyoming launches first state-backed stablecoin…
Eight Sleep (TBPN sponsor) raises $100M Series D from HSG, Valor, Founders Fund, and more…
Timeline
