E44: USA's Afghanistan embarrassment, China's new algo laws, future of robots + Italy recap!

Multiple independent sources state that Callin’s public launch occurred on September 2, 2021. David Sacks’ own launch blog post on Medium, dated September 2, 2021, explicitly says “Today is the public launch of Callin,” confirming that this was the day the app became publicly available on the Apple App Store. (medium.com) TechCrunch’s coverage likewise reports that Callin “launches” on September 2, 2021, alongside its $12M Series A announcement. (techcrunch.com) Forbes also runs a feature on the same date describing Sacks as “launching the ‘social podcasting’ platform Callin on the Apple App Store.” (forbes.com) Additional funding and news databases (Crunchbase, VC News Daily) list the product launch and funding announcement on September 2, 2021. (crunchbase.com) Together, these sources confirm that the Callin app’s public launch did indeed occur on 2021-09-02, matching the prediction.

Sacks predicted that, in the years after the 2021 U.S. withdrawal, China would cut deals with the Taliban to build a highway (major transport infrastructure) inside Afghanistan specifically to reach, extract, and export Afghan rare earth and other minerals to feed China’s economy.

What has happened so far:

1. China–Taliban resource deals did materialize, but mainly as mining/oil contracts, not transport infrastructure. In January 2023, the Chinese company CAPEIC (a CNPC subsidiary) signed a long-term, hundreds‑of‑millions‑dollar oil extraction deal in the Amu Darya basin with the Taliban government, and Chinese firms have been negotiating over the Mes Aynak copper mine and other minerals.(en.wikipedia.org) A 2025 analysis notes that Chinese firms now control a large share of active extraction projects under Taliban “security‑for‑minerals” arrangements, especially for rare earths and lithium—but this still concerns mining operations, not new roads or railways.(rareearthexchanges.com)

2. Afghanistan remains only marginally integrated into China’s Belt and Road transport network. A survey of Afghanistan–China relations notes that, as of 2023, Afghanistan’s security situation prevented it from becoming a major part of the BRI.(en.wikipedia.org) While there are long‑standing proposed rail corridors (e.g., Five Nations Railway Corridor, Trans‑Afghan Railway), they have moved slowly for years and are not yet built, let alone operating to evacuate Afghan rare earths.

3. Post‑2021, China and the Taliban have agreed only at a high level to extend existing corridors, not to build the kind of dedicated mineral‑export “superhighway” Sacks described. In May 2025, China, Pakistan, and the Taliban government announced that the China–Pakistan Economic Corridor (CPEC) would be extended into Afghanistan and that CPEC would be linked to the proposed Trans‑Afghan Railway.(en.didpress.com) These are framework political agreements about future connectivity; there is no evidence of concrete, financed Chinese road or rail projects inside Afghanistan that are being built or contracted specifically to move Afghan rare earth/mineral output to China.

4. Key reporting indicates China is avoiding large infrastructure builds in Afghanistan, despite mining interest. A 2025 Washington Post report on the Taliban’s attempted highway through the Wakhan Corridor—envisioned as a route to China—notes that the Taliban are struggling to pave a short stretch with their own limited funds and explicitly states that, despite Taliban claims, China has “so far stayed away from undertaking large infrastructure projects” in Afghanistan.(washingtonpost.com) More generally, prior analyses of China’s role in Afghanistan stress that, beyond a few modest projects, Beijing has refrained from new large‑scale infrastructure there, even after signing a BRI MoU, due to security and political risks.(ecfr.eu)

Taken together, the core of Sacks’s prediction—that China would, in the near term, strike deals with the Taliban to build a major new highway/transport corridor inside Afghanistan whose primary purpose is to unlock and export Afghan rare earth and mineral wealth to China—has not occurred as of late 2025. China has signed notable mining and oil contracts and has talked about future corridor extensions, but it has not committed to or begun constructing the kind of dedicated, large‑scale Afghan transport infrastructure he described. Therefore, the prediction is best assessed as wrong.

As of November 30, 2025, Taiwan remains self-governed under the Republic of China (ROC) with its own democratic institutions, military, and de facto autonomy. The People’s Republic of China (PRC) claims Taiwan but has not brought it under full PRC/CCP control, nor executed a successful invasion or forced unification.

Key observable facts:

1. No PRC takeover of Taiwan yet

   • Taiwan continues to hold its own elections, operate its own government, and field an independent military.
   • There has been no formal PRC annexation, forced unification, or regime change in Taipei.

2. No US–PRC war over Taiwan to date

   • While US–China tensions over Taiwan have significantly increased (military exercises, arms sales, high-level visits, and repeated PLA air and naval activity around Taiwan), there has been no full-scale PRC attack on Taiwan that would test whether the US would “really engage” militarily in the way described in the prediction.

Chamath’s prediction has two linked components:

1. PRC eventually brings Taiwan under complete Chinese (PRC/CCP) control.
2. When that happens, the US ultimately chooses not to intervene militarily in a meaningful way.

Because the first event (full PRC control over Taiwan) has not occurred at all yet, the second (US non-intervention) has not been tested. The prediction is about an open‑ended future scenario without a specified deadline, so it cannot yet be judged as right or wrong—only that it has not (yet) come to pass.

Therefore, the correct classification as of now is inconclusive (too early).

As of November 30, 2025, many large Chinese companies that use VIE-based structures remain listed on U.S. exchanges via ADRs, so the blanket claim that such ADRs would have to be delisted has not come true.

Examples:

• Alibaba Group (BABA) still trades on the NYSE and uses a Cayman Islands VIE structure to allow foreign investors exposure to its Chinese operations.
• JD.com (JD) continues to be listed on NASDAQ, also via a VIE-based offshore entity.
• Baidu (BIDU) is still traded on NASDAQ through a similar VIE/ADR structure.
• Numerous other Chinese tech and internet companies with VIE-based ADRs continue to be listed; some have added secondary or dual-primary listings in Hong Kong but have not delisted from U.S. markets.

While there have been notable individual delistings or withdrawals (e.g., DiDi Global and some smaller firms) driven by Chinese regulatory pressure and the U.S. Holding Foreign Companies Accountable Act, the prediction was categorical: “You’re going to have to delist these ADRs.” Interpreted as a broad outcome for Chinese VIE-based ADRs in general, this has clearly not happened by late 2025. Many remain active and liquid on U.S. exchanges.

Therefore, the prediction is wrong in its broad, deterministic form.

Available evidence indicates that China did not move from its 2021–2022 tech and data regulatory push to cancel the VIE structures underpinning major Chinese internet/tech platforms; instead, it formally legalized and regulated them.

1. Regulators explicitly chose to allow VIEs rather than cancel them.

   • In late 2021 and 2023, the China Securities Regulatory Commission (CSRC) issued draft and then final rules on overseas listings. These rules do not ban VIEs; rather, they put VIE-structured overseas listings under a filing regime and state that companies with VIE structures that meet compliance requirements are eligible to list overseas after filing.(cnbc.com)
   • Law-firm and regulatory summaries note that VIEs are now expressly included in the CSRC filing system and are “no longer a grey-area scheme,” signalling regulatory acceptance rather than cancellation.(investmentlawwatch.com)

2. Major tech companies using VIEs remain listed and operating via those structures.

   • The largest Chinese tech firms that rely on VIE structures for overseas listings—such as Alibaba, JD.com, Baidu and others—remain listed on U.S. and/or Hong Kong exchanges as of 2025. There has been political pressure in the U.S. to delist some of them on national security grounds, but that is a U.S. capital-markets issue, not China canceling their VIEs.(cnbc.com)
   • The new Chinese rules apply prospectively via filings and compliance reviews; they do not retroactively invalidate or force unwinding of the core VIE structures for these major tech platforms.(mayerbrown.com)

3. Where VIE-related prohibitions did occur, they were sector-specific (education), not a broad tech‑platform move.

   • In July 2021, China’s "Double Reduction"/education crackdown barred foreign capital from controlling or participating in curriculum-based tutoring businesses via VIE arrangements, and required any violations to be rectified.(sec.gov)
   • This effectively killed the VIE-based, for‑profit K‑12 tutoring model (e.g., New Oriental, TAL), but that is a targeted ban in the education sector, not the broader "most important" tech platforms that were the focus of investors’ VIE concerns (internet platforms like Alibaba, Tencent, Meituan, etc.).

Because the Chinese government’s post‑2021 regulatory trajectory normalized and regulated VIEs for major tech companies instead of canceling them, the core forecast that new tech/data rules were a prelude to canceling those VIE structures has not materialized. The narrow education‑sector VIE ban is too limited and sector‑specific to count as the predicted broader move against major tech VIEs.

On balance, this prediction is wrong.

Current scientific and public‑health consensus is that SARS‑CoV‑2 has become an endemic virus that continues to circulate globally, much like seasonal influenza and the common cold, and will persist for the foreseeable future rather than being eradicated.

• The WHO and many national health agencies now describe COVID‑19 as being in an endemic phase in most regions, meaning the virus continues to circulate with predictable or recurring patterns instead of being eliminated.
• No credible body expects eradication; instead, long‑term management strategies (vaccination updates, treatment protocols, and surveillance) are framed around ongoing, indefinite circulation.

Given that by late 2025 COVID‑19 is clearly still present worldwide and treated as a permanent part of the infectious‑disease landscape, the prediction that “COVID's going to be around forever, like the cold or the flu” has been borne out in practice.

Chamath’s claim has two parts. For SpaceX’s Starlink, it is clearly correct: by 2025 Starlink has launched over 7,600 satellites (about 65% of all active satellites), is available in more than 140 countries, and has grown from 1 million subscribers in 2022 to about 8 million by November 2025, with multi‑billion‑dollar annual revenues and a central role in SpaceX’s valuation—i.e., it is a commercially significant, widely deployed service.(en.wikipedia.org) For Tesla’s humanoid robot (Optimus/Tesla Bot), however, the situation is different: as of late 2025, Optimus units are doing limited tasks inside Tesla facilities and public demos, and Musk repeatedly talks about thousands of internal robots in 2025 and large‑scale production/sales starting in 2026, but major business and finance outlets note that Optimus robots are still not yet on the market or sold commercially, with scale production targeted for 2026 and beyond.(cnbc.com) So today the Starlink half of the prediction has clearly come true, while the Tesla Bot half has not yet matured into a real external product line (though it might in the future). Because one component is fulfilled and the other is still pending rather than clearly failed, the overall, combined prediction is best rated as inconclusive at this time.

By late 2025, parts of the prediction are directionally supported, but the key elements have not yet occurred, and the relevant time horizon (Mars/space use, large internal fleets) clearly extends beyond today.

What has happened so far

• Tesla has indeed developed its humanoid robot, Optimus, to a working prototype stage and begun limited deployment inside its own facilities. Reports indicate over 100 Optimus units operating in Tesla factories on repetitive tasks, plus additional units in offices and for demo/service roles (e.g., Tesla Diner, public events).(opentools.ai)
• Optimus is still not a mature, fully autonomous product: investigative reporting describes it as heavily reliant on human training data, with lab setups where workers demonstrate motions; the robot still stumbles and often needs support rigs.(businessinsider.com)
• Tesla has no external commercial customers yet. Musk and Tesla repeatedly say Optimus will be used internally first, with sales to other companies planned only after Tesla has deployed and refined robots in its own factories (targeting 2026 and beyond).(xatakaon.com)

These facts are consistent with the spirit of the podcast claim that Elon/Tesla would be the robot’s first major customer, but they are still early, small‑scale pilots rather than "the first 1,000" or "thousands" of units.

What has not happened yet

• The "on the order of thousands" deployment is still a goal, not reality. Musk and analysts repeatedly project 1,000–2,000+ Optimus units in Tesla factories by the end of 2025 and tens of thousands per year in 2026+, but these numbers are planning targets; public reporting through late 2025 still describes only limited internal use and low-volume production, with full-scale production pushed back toward 2026.(investor.wedbush.com)
• There is no evidence that Optimus is working in Boring Company tunnels. The Boring Company’s automation push is focused on increasingly autonomous tunnel-boring machines and “Zero-People-in-Tunnel” continuous mining, not humanoid robots hauling rocks.(urca.foundation)
• Optimus has not yet flown to space or been used for space-station or Mars construction. Musk has announced plans for Starship missions to Mars in 2026 that would carry Optimus robots, but these are future missions that have not yet occurred.(investing.com)

Why the verdict is “inconclusive”

• The core structural claim—Elon Musk, via his companies, will be the initial major customer for Tesla’s humanoid robot—is supported so far: as of 2025, all known Optimus deployments are internal to Tesla, and external customers are planned only later.
• However, the specifics Jason gave (first ~1,000 robots used across Tesla factories, Boring Company tunnels, and Mars/space-station construction) depend on events that are explicitly planned for the late 2020s and beyond, especially Mars work. Those milestones have neither been achieved nor clearly falsified by 2025.

Because the prediction’s key conditions (thousands of units across Musk companies, robots used in Boring tunnels and in space/Mars construction) lie in a timeframe that has not yet arrived, we cannot judge it as clearly right or wrong as of November 30, 2025. Hence the status is "inconclusive (too early)".