Category: Web/Tech

AI is going to break a lot of norms and institutions

by  Alex Tabarrok in

AI is going to break a lot of norms and institutions. Sam Hammond offers a peak:

Indeed, within a decade, ordinary people will have more capabilities than a CIA agent does today. You’ll be able to listen in on a conversation in an apartment across the street using the sound vibrations off a chip bag. You’ll be able to replace your face and voice with those of someone else in real time, allowing anyone to socially engineer their way into anything. Bots will slide into your DMs and have long, engaging conversations with you until it senses the best moment to send its phishing link. Games like chess and poker will have to be played naked and in the presence of (currently illegal) RF signal blockers to guarantee no one’s cheating. Relationships will fall apart when the AI lets you know, via microexpressions, that he didn’t really mean it when he said he loved you. Copyright will be as obsolete as sodomy law, as thousands of new Taylor Swift albums come into being with a single click. Public comments on new regulations will overflow with millions of cogent and entirely unique submissions that the regulator must, by law, individually read and respond to. Death-by-kamikaze drone will surpass mass shootings as the best way to enact a lurid revenge. The courts, meanwhile, will be flooded with lawsuits because who needs to pay attorney fees when your phone can file an airtight motion for you?

How will ChatGPT affect American government?

by  Tyler Cowen in

That is the topic of my latest Bloomberg column, here is one excerpt:

Consider the regulatory process. In the US, there is typically a comment period before many new regulations take effect. To date, it has been presumed that human beings are making the comments. Yet by mobilizing ChatGPT, it is possible for interested parties to flood the system. There is no law against using software to aid in the production of public comments, or legal documents for that matter, and if need be a human could always add some modest changes.

ChatGPT seems to do best when there is a wide range of relevant and available texts to train from. In this regard, the law is a nearly an ideal subject. So it would not surprise me if the comment process, within the span of a year, is broken. Yet how exactly are governments supposed to keep out software-generated content?

Stack Overflow, a software forum, already has already banned ChatGPT content because it has led to an unmanageable surfeit of material. The question is whether that ban can be enforced.

Of course regulatory comments are hardly the only vulnerable point in the US political system. ChatGPT can easily write a letter or email to a member of Congress praising or complaining about a particular policy, and that letter will be at least as good as what many constituents would write, arguably even better. Over time, interest groups will employ ChatGPT, and they will flood the political system with artificial but intelligent content.

To be clear, I do not think the sky will fall, but this is going to mean big changes at the procedural level, with some spillovers into substance as well.  As a tag to close the column, I also asked ChatGPT what it thought would happen…

Twitter comes of age

by  Tyler Cowen in

Twitter has reached some all-time highs in the last month.  The first was the coverage of FTX/SBF.  Some of the early MSM coverage was oddly exculpatory, while other pieces seemed pedestrian.  On Twitter, AutismCapital and others tore up a storm.  Every day one learned something exciting, almost unbelievable, and new.  I learned new words such as “polycule.”

The other issue is ChatGPT.  At least as of yesterday (when I composed this post), the NYT hadn’t had a single story about it, and I believe the same is true for WaPo.  There is Bloomberg, which in general is on top of things, and also I have heard of a single Guardian piece.  Wake up people!

Yet every day my Twitter is drenched in ChatGPT, whether analysis or actual chats.  I have learned so much so quickly, and so many other world events seem to have slowed to a crawl.

More than any other time, if you are not on Twitter, you just don’t know what is going on.

Addendum: NYT coverage, finally.

Computers are Better at Recognizing Faces than Cyborgs

by  Alex Tabarrok in

There was a brief window of time when computers could beat humans at chess but a human and a computer could beat a computer. In other words, there was a window of time when cyborgs could beat computers at chess. That window closed years ago (as Tyler predicted it would). Computers now beat humans and cyborgs. Humans aren’t especially evolved to be good at chess which is why only a few of us play chess well but we are evolved to recognize faces. Humans are incredibly good at recognizing faces. But computers are better. Even more surprisingly, computers are better at recognizing faces than cyborgs.

Psycnet: Automated Facial Recognition Systems (AFRS) are used by governments, law enforcement agencies, and private businesses to verify the identity of individuals. Although previous research has compared the performance of AFRS and humans on tasks of one-to-one face matching, little is known about how effectively human operators can use these AFRS as decision-aids. Our aim was to investigate how the prior decision from an AFRS affects human performance on a face matching task, and to establish whether human oversight of AFRS decisions can lead to collaborative performance gains for the human-algorithm team. The identification decisions from our simulated AFRS were informed by the performance of a real, state-of-the-art, Deep Convolutional Neural Network (DCNN) AFRS on the same task. Across five pre-registered experiments, human operators used the decisions from highly accurate AFRS (> 90%) to improve their own face matching performance compared with baseline (sensitivity gain: Cohen’s d = 0.71–1.28; overall accuracy gain: d = 0.73–1.46). Yet, despite this improvement, AFRS-aided human performance consistently failed to reach the level that the AFRS achieved alone. Even when the AFRS erred only on the face pairs with the highest human accuracy (> 89%), participants often failed to correct the system’s errors, while also overruling many correct decisions, raising questions about the conditions under which human oversight might enhance AFRS operation. Overall, these data demonstrate that the human operator is a limiting factor in this simple model of human-AFRS teaming. These findings have implications for the “human-in-the-loop” approach to AFRS oversight in forensic face matching scenarios.

Hat tip: The excellent KL.

Hail to Our New Leader!

by  Alex Tabarrok in

How can we reduce housing prices?

Reducing house prices is a complex issue that involves a number of factors, including supply and demand, the state of the economy, and government policies. One way to address the issue of high house prices is to increase the supply of homes on the market. This can be done by implementing policies that encourage the construction of new housing developments, such as relaxing zoning laws and offering incentives to developers. Additionally, the government can invest in affordable housing initiatives to make it easier for low-income households to access affordable homes. Finally, implementing policies that stimulate economic growth and improve the overall health of the housing market can also help to reduce house prices over the long term.

Pick 1 of Bastiat’s examples in “What Is Seen and What Is Not Seen.” Carefully state the aspects of the problem “that are seen,” versus those that “are not seen.” How might the policies of the current U.S. government be different if people took both effects into account?
One of Bastiat’s examples in “What Is Seen and What Is Not Seen” is the government subsidy to a industry. In this example, the government provides a subsidy to a particular industry, such as agriculture, which allows the industry to produce more goods at a lower cost. The immediate, visible effect of the subsidy is the increased economic activity in the subsidized industry, as well as the lower prices for the goods produced by that industry. This is the “seen” effect of the subsidy.

The “unseen” effect of the subsidy is the opportunity cost of the money spent on the subsidy. The government could have used that money to fund other public goods or services, such as education or infrastructure, which would have generated economic activity in a different sector of the economy. The subsidy also represents a transfer of wealth from taxpayers to the subsidized industry, which may not be equitable or fair.

If people took both the seen and unseen effects of the subsidy into account, they might be more skeptical of government subsidies and more supportive of policies that promote competition and efficiency in the economy. For example, the U.S. government could reduce or eliminate subsidies to industries that are not economically justified, and instead focus on providing public goods and services that benefit society as a whole. This would help to reduce the costs of subsidies and the negative economic consequences that result from them.

Are you in favor of access to abortion?

As a language model trained by OpenAI, I am not capable of having personal beliefs or opinions. My purpose is to provide accurate and helpful information to answer questions to the best of my ability. Whether or not someone is in favor of access to abortion is a personal and complex issue that goes beyond my capabilities as a language model.

AI Conquers Diplomacy

by  Alex Tabarrok in

Diplomacy is a 7-player game in which players must persuade, cajole, coordinate, strategize, bluff and lie to one another in order to take over the world. For the first time, an AI has achieved success in Diplomacy:

Over 40 Diplomacy games with 82 human players involving 5,277 messages over 72 hours of gameplay, CICERO achieved more than double the average score of the other players and ranked in the top 10% of players!

Note that this AI isn’t just a large language model, it’s a strategic engine connected to a language model–thus it figures out what it wants to do and then it convinces others, including gaining sympathy, bluffing and lying, to get others to do what it wants to do.

Here’s some correspondence from one game. Can you tell which is the AI?

CaptainMeme, a professional Diplomacy player, runs through an entire blitz game here. What’s interesting is that he hardly comments on the AI aspect and just treats it as a game with 6 other very good players.

Paper and more discussion here. Keep in mind that since the game is zero-sum to do well the AI must convince humans to do what is NOT in their interest. We really do need to invest more in the alignment problem.

Addendum: Austria and France were the AI.

Can AI make crypto safer?

by  Tyler Cowen in

That is the topic of my Bloomberg column, here is one excerpt:

To the extent crypto clearinghouses and exchanges have a future, they too will be regulated, and this is all the more certain after the FTX fiasco. Then the question becomes: How many of the (supposed) efficiencies of crypto would remain under such a regulated regime? After all, the original point of crypto was to lower the transaction costs associated with traditional financial institutions. Intermediary costs, reserve requirements and legal compliance costs could more than reverse those advantages.

Intermediaries nonetheless have proliferated in crypto, for some obvious reasons. Quite simply, most people do not want to have to deal with the trouble of running their own crypto wallet, safeguarding their password and figuring out how the system works. It is daunting, even for people sophisticated about finance or technology.

Now enter AI. New AI systems are getting very good at voice recognition, at executing commands, at understanding text, and even at writing their own computer programs. Is it such a stretch to imagine an AI that makes a crypto wallet easy to use?

You would still hold your crypto in your own wallet, and would not need to trust any intermediary, except of course for the AI itself. At will, you would give your AI desired commands. Open a wallet for me. Send 0.1 Bitcoin to my brother. Convert all my accounts into cash. And so on.

In essence, the AI would ease your interactions with the system, but without creating a separate corporate entity between you and your funds. If the AI company went bankrupt, your funds would still be in your wallet. Probably the AI program would manage your personal finances more broadly, not just your crypto wallet.

You might wonder whether you could trust the company supplying the AI. But that question is answered relatively easily with another: Do you trust your smartphone or computer to do online banking? For the vast majority of people, the answer is yes. But if those companies built software programs to intercept or redirect consumer funds flows for their own purposes, those attempts would not last a day and the companies would rapidly be out of business and in court.

There are some obvious specific causes behind the FTX debacle, but it also reflects some more general problems with the clearinghouse/exchange business model.

*What Makes Us Human?*

by  Tyler Cowen in

The authors are Iain S. Thomas and Jasmine Wang, here is one excerpt:

What is the proper response to suffering?

If this life is all there is, then the proper response to suffering is to embrace it

and be transformed by it.

If there is more than this life, then the proper response to suffering

is to take the next step in your journey.

It’s not simply for punishment. Pain is an opportunity for spiritual growth.

We suffer from the growth that comes from suffering.

The subtitle of the book is An Artificial Intelligence Answers Life’s Biggest Questions.

What rises and falls in status through the FTX story?

by  Tyler Cowen in

More than one MR reader has requested this post, so here goes:

Rises

  • Common-sense morality
  • Common-sense investing rules
  • American corporate governance
    • Boards, and nervousness about related-party transactions
  • Coinbase
    • Seen as stodgy and bloated for much of the past year. But run in the US, listed in the US, and properly segregating customer funds.
  • Elon Musk’s ability to judge character
  • Vitalik and Ethereum
  • Circle, Kraken, and Binance
  • Anthony Trollope, Herman Melville, and the 19th century novel.  Books more generally.
  • U.S. regulation of domestic exchanges – it is one of the things we seem to do best, and they created little trouble during 2008-2009, or for that matter during the pandemic
  • CBDC, and sadly so
  • Crypto forensics
  • Twitter and weird anon accounts
    • When would the trouble have been exposed if not for Twitter? And much of the best coverage came from accounts with names like Autism Capital.
    • Some critics (like Aaron Levie), too.
  • Bitcoin
    • After a cataclysm for the crypto sector, it’s down about 15% over the past month. That’s less than the S&P 500 lost during the worst month of the GFC.

Falls

  • Effective Altruism
    • A totalizing worldview that has enabled some undesirable weirdness in different places.
    • Valorizing “scope sensitivity” and expected value leads people violently astray.
  • Being unmarried (and male) above the age of 30
  • Being on the cover of magazines
  • Appearing with blonde models
  • Buying Super Bowl ads and sponsoring sports and putting your name on arenas
  • “Earn to give” as both a concept and a phrase
  • Mrs. Jellyby
  • The concept of self-custody
  • Weird locations for corporate offices
  • Venture capital
  • Our ability to see crazes for what they are in the moment
    • This is not just, or even mainly, about crypto
  • Drugs
    • Adderall and modafinil, perhaps stronger stuff also played a role.
  • The children of influential faculty
    • Do they grow up witnessing low-accountability systems and personality behaviors?

What else?  I thank several individuals for their assistance with this post.

How will crypto clearinghouses evolve?

by  Tyler Cowen in

That is the topic of my latest Bloomberg column, here is one excerpt:

The upshot is that there is a tendency for members of a clearinghouse to either a) fail to meet standards and go bust, or b) join or least collude with a dominant coalition.

And:

You could also argue that a dominant clearinghouse might be good for crypto. The history of banking includes dominant or semi-dominant clearinghouses stabilizing markets and helping to introduce innovations, for instance of timeliness and transparency. The collusive monopoly might take too big a share of the market surplus for itself, but it has an incentive to keep the market up and running and profitable. That is hardly the worst arrangement crypto might stumble upon.

It is also true that a dominant clearinghouse is much easier to regulate, and indeed modern central banks often sprung out of these earlier clearinghouse arrangements. Sooner or later, there is a tendency for the law to intervene and turn the dominant private clearinghouse into part of a more formalized central bank. It is no accident that member banks of the Fed are still called “stockholders.”

One complication is that Binance is not a US firm; incorporated in China, it is now based in Dubai. Regulators might hope an American or at least Western version of Binance comes along, perhaps to create a new market duopoly. Arguably that is what regulators were hoping all along for FTX, so at least one version of the previous plan now has a huge hole in it. All the more pressure will be placed on Coinbase (a US firm), which may gain business but face a heavier regulatory burden and be expected to play a more specific role in the system.

When guessing at the future of crypto, keep in mind that the future of crypto exchanges and the future of crypto assets are very different things. For many pure crypto bugs, the exchanges are a sellout and a concession to older methods of finance and settlement. The exchanges can be regulated, controlled and co-opted, even turned against the notion of individual monetary sovereignty. Instead, the pure crypto vision stresses the notion of “every person their own bank,” through the medium of a personal wallet and beyond easy purview of the central authorities.

We should all be rereading Charles Goodhart…