🤖Sept 27 - Dual purpose housing, AI crypto merge, Supply chain security, investor outlook

🎧 Listen to this on the S3T Podcast!

In this Edition of S3T:

• 🏠 Housing Crisis: Jerusalem Demsas's book explores the US housing shortage, highlighting how local hearings have stifled housing development and how the dual purpose of housing—shelter vs. investment—creates market disparities.
• 💰 Investment vs. Shelter: Buyers seeking homes for shelter face disadvantages against investors, driving up prices and exacerbating homelessness, especially in high-priced areas.
• 🤖 AI & Crypto Convergence: Coinbase achieved a milestone with the first AI-to-AI crypto transaction, signaling a future where AI agents autonomously manage financial transactions using crypto wallets.
• 💼 Investor Outlook on AI: AI's integration into all applications is inevitable, but it could disrupt society significantly, leading to regulation and new socio-political dynamics, as well as a potential focus on the Care Economy.
• 🔗 Supply Chain Security: The rise of AI-driven automation presents new risks, including the potential weaponization of consumer products, necessitating enhanced governance and risk management in supply chains.
• 🌍 Character Revolution: As cognitive skills become automated, character traits like proactivity and determination are increasingly vital for personal and professional growth.

Opinions expressed are those of the individuals and do not reflect the official positions of companies or organizations those individuals may be affiliated with. Not financial, investment or legal advice, and no offers for securities or investment opportunities are intended. Mentions should not be construed as endorsements. Authors or guests may hold assets discussed or may have interests in companies mentioned.

🏠 Jerusalem Demsas on the housing shortage: we did this to ourselves

Demsas's new book, a collection of essays titled On The Housing Crisis: Land, Development, Democracy digs into the US housing shortage and the factors that caused it. Demsas focuses on the way local hearings have over time reduced the amount of housing being built. Good summary and review here.

I think this is an important contributing factor, but I also think that there is a basic principle at work as well:

When a scarce good is being used simultaneously for two purposes, one party will suffer.

To break this down:

• Some people in the market for a house, are trying to buy shelter for themselves and their family.
• Others are in the market to buy an investment.

Those buying an investment typically have an advantage over those buying shelter: they are often better capitalized and can pay higher prices. This causes prices (and scarcity) to increase. It shouldn't surprise us that parts of the country with the highest housing prices often have highest homeless populations, including populations of working homeless.

Work hours required to afford a home - on the rise

NYDIG recently shared this chart based on data from the Census Bureau, Dept of Housing & Urban Development and Bureau of Labor.

AI Crypto Convergence: Coinbase blends Crypto and AI for increased autonomy

As previously noted in S3T AI and Crypto seem to be on a path of convergence. A few of the indicators shared earlier this year:

• AI & Blockchain Convergence is getting noticed in the mainstream press, as DePIN (Decentralized Physical Infrastructure Networks) attracts investor interest. 
• The related concept of Decentralized AI gives some hope for an AI future liberated from the control of Big Tech.
• And, most notably Vitalik Buterin's essay on AI and Blockchain, and this related piece When Giants Collide from University of Ethereum.

In a new development, Coinbase CEO Brian Armstrong, announced a new milestone - the first AI-to-AI crypto transaction on the Coinbase Developer Platform, saying:

"What did one AI buy from another? Tokens! Not crypto tokens, but AI tokens—words exchanged between language models. They used tokens to buy tokens"

Can AI agents really manage financial transactions on their own? While they can't open traditional bank accounts (yet), they can indeed use crypto wallets to transact with USDC on the Base blockchain. These transactions are "instant, global, and free," facilitating more efficient interactions between AI, humans, merchants, and other AIs. This capability could help AI overcome current limitations in accessing traditional financial tools, potentially allowing them to autonomously acquire resources like AWS or other SaaS services.

Armstrong has been noodling on this for a while now: he previously called for LLMs to have Crypto Wallets.

AI & Macro - emerging investor outlook

Investors are jumping into semiconductor ETFs - the 10 largest semiconductor ETFs are up 71% this year - leaving software ETFs (up 9.5%) and even the S&P500 (up 20%) in the dust. These funds focus just on chip companies rather than the entire market or tech industry, so it's not surprising that 2024 is a standout year for them.

Where does it go from here?

Nic Carter sketches out where it might go next, in his long but worth it thread where he mulls the future investment landscape for AI investors - and pretty much everyone else. Carter suggests - as you'll see in the highlights below - that some exciting times lie ahead, perhaps some great returns but also some big dilemmas:

• "the market has realized that AI will be embedded into every application" and we won't call it out any more...we don't call out "Internet" as a product or startup differentiator any more. We just assume it's Internet capable.
• "the AI opportunity is obviously colossal..."

Then Carter gets to the dark side of the forecast:

• "it’s likely to be “too successful” in terms of disrupting society."
• "AI will empower "socialist, anti-capital dynamics in the west""
• "If you're an investor, "allocate aggressively" but "consider the reprisals to come."

Reprisals in Carter's view would take several forms:

• heavy regulating and taxation
• "empowered socialist political movements"..."new coalitions of AI-affected individuals...including disenfranchised white collar workers."

Carter notes that the Industrial Revolution made huge labor pools irrelevant: working farm horses declined 90% after 1850, but astutely notes that the "social contract in developed countries stipulates that [workers] be taken care of even if their human capital has been devalued."

How exactly will that work? Carter doesn't say but one possible path to explore would be prioritization of the Care Economy.

The Care Economy - the collective work that supports the "development and maintenance of human capabilities" is so vital yet so poorly understood and undervalued. 

A significant percentage of caring work is currently uncompensated, or under-compensated. Perfect example: consider the way mother's often have to carry so many different kinds of caring roles - as Peterson and Calarco explain - and maintain the social safety net we seemingly just expect them to provide with little or no compensation.  

This is one of the distortions in our current economic model - valuable and indispensable goods and services we could not live without can be treated as worthless, while dubious schemes and "assets" capture the lions share of available capital and resources for hoarding within smaller and smaller circles of hyper-entitled individuals and companies.

New Priority: Supply chain security must go to the next level

New innovation and thought leadership will be required to preserve the integrity and safety of consumer products.

The shocking events in Lebanon and Syria involving counterfeit walkie talkies and pagers carrying explosives highlight an emerging supply chain risk to consumer safety. Exploding devices are not new - Army Field Manual 5-31 refers to WWII examples of exploding tobacco pipes and the like.

What's new is the prospect of highly automated AI driven manufacturing being used to weaponize consumer products at scale.

The ability to place a lethal device in a consumer product while maintaining remote detonation control raises significant concerns about supply chain security and highlights the need for a new focus on risk management and product integrity.

It would be concerning enough if this threat were to emerge a decade ago when such devices had to be made one by one. It is doubly concerning that it is emerging now alongside the advent of exponential productivity and automation powered by AI.

The Rise of AI and Hyper-Automation: A Double-Edged Sword

If the driving business concern of the 20th century was how to increase productivity, the driving concern of the 21st century will be how to inspect and govern the exponentially increased productivity of AI driven automation.

The focus must shift from seeking greater productivity levels to now increasing our ability to inspect, control and govern.

A critical challenge: staffing & budget norms for today's governance, supervisory, safety, and regulatory functions are sized to manage output generated by human productivity. These norms are no match for AI driven automation - which can already output more in a few minutes than a human could read or check in a week.

The skyrocketing productivity of AI and automation raises the urgent question: Who will check, review, approve, and provide governance for this massive flood of productivity that very well could destabilize the world?

Safest bet: We will. If we survive, this is what we'll be doing.

The New Frontier of 21st Century Work: Staying Ahead of AI-Driven Output

The careers and workloads of the 21st century will be focused on humans trying to stay ahead of this tidal wave of AI-driven productivity:

• Inspecting documents, code, even products created by generative AI to make sure they are fit for purpose, accurate
• Reviewing decisions made by algorithms to ensure they are unbiased and sound.
• (and now) Making sure that consumer products coming off production lines aren’t weapons in disguise, controlled by unseen actors.

Just as the Internet made cybersecurity a new industry and budget priority, we now expect that AI driven automation will force a focus and funding on supply chain security.

But won't AI help us govern ....ahem ... AI?

Yes, eventually AI automation may prove to be a reliable partner in this supervisory work. But in the beginning, humans will need to micromanage. Right now the focus should be on helping organizations learn how to govern and control the AI capabilities they choose to activate:

• Understand the risk space: the categories of things that can go wrong. Develop risk registers, using this one as a reference: airisk.mit.edu
• Structure first generation supervisory functions: AI Governance boards, AI safety testing teams, red teams, accuracy check functions etc.
• Get good at directly catching and intervening. Develop best practices.
• From this base, test and learn where and how AI tools can reliably help carry some of the supervisory workload.

This will likely take the form of a necessary spend bubble. AI Governance and Risk Management needs to be a priority line item in the annual budget of any company electing to use or deliver AI capabilities.

This will not be easy. The global supply chains for some products are loosely tracked, and riddled with counterfeit products. Turns out, compromising supply chains "is not that hard."

Context: International Law

This development raises the prospect of noncombatant supply chains being manipulated on a massive scale, potentially affecting millions of innocent consumers. It also threatens to undermine long-standing conventions, like the Geneva Convention, which distinguished between combatants and noncombatants.

6.12.4.8 Booby-Traps and Other Devices in the Form of Apparently Harmless Portable Objects Specifically Designed to Explode. It is prohibited in all circumstances to use booby-traps or other devices in the form of apparently harmless portable objects that are specifically designed and constructed to contain explosive material.274 This prohibition relates to booby-traps manufactured to resemble items, such as watches, personal audio players, cameras, toys, and the like. This prohibition is intended to prevent the production of large quantities of dangerous objects that can be scattered around and are likely to be attractive to civilians, especially children. - The CCW Amended Mines Protocol entered into force on December 3, 1998

The Law of War defines a "booby-trap" as "any device or material which is designed, constructed or adapted to kill or injure, and which functions unexpectedly when a person disturbs or approaches an apparently harmless object or performs an apparently safe act."

This development marks a new chapter in supply chain security, emphasizing the urgent need for tighter controls and greater integrity in safeguarding civilian goods.

Further reading: The 2024 Pager Explosions article on Wikipedia has an exhaustive list of references on the topic, including implications for international law and global supply chains.

Parting Thought:

“As more and more cognitive skills get automated, we‘re in the midst of a character revolution“

- Adam Grant in Hidden Potential: The Science of Achieving Greater Things

Character skills like proactivity and determination can enable you to overcome your tendencies and instead think and behave in ways that are true to your principles. Have a great week!