DRAFT: Feeding Audio to Chat-GPT3 & Getting Reasonable Results; & BRIEFLY NOTED: For 2023-02-28 Tu

Getting something useful out of Chat-GPT3; book as experience machines; Matt Levine snarks; Tom Nichols mourns...

Getting something useful out of Chat-GPT3; book as experience machines; Matt Levine snarks; Tom Nichols mourns...

---

FOCUS: Feeding the Audio from Hexapodia to Chat-GPT3:

Taking:

PODCAST: “Hexapodia” is þe Key Insight! XLVII: “Polycrisis” Was Just the New Cold War All the Time!

And feeding the audio to MacWhisper and then to Chat-GPT3 with this prompt:

The text below is the noisy text of a podcast discussion between Noah Smith and Brad DeLong. Please clean the text up and summarize it, making it grammatical and focusing on the major ideas:

What will it produce? Will it produce anything useful? It produces this:

In this podcast, Brad DeLong and Noah Smith… start by discussing a quote[s] from Oscar Wilde and Kurt Cobain about individualit,y and the dangers of wishing to be someone else. They then shift the conversation to the topic of heroin, questioning how a plant could have such an intense effect on the human brain. They conclude that the refinement of opium through smoking and distillation is what makes it so potent, and that the evolution of these plants did not account for human refinement of the drug. They also note… caffeine is an example of another widely-used substance that has mind-altering effects. Brad and Noah [then] discuss the disturbing reality of corporations intentionally addicting people to drugs like heroin in order to increase profits. They both express their disappointment and frustration with this unethical behavior and its devastating impact on individuals and society as a whole.

Brad then changes the subject to the current state of the world. He asks Noah if he has any better news to share. The two briefly discuss the topic of good news before moving on to the next topic, which is the ongoing Cold War between the US and China and the potential impact on global trade. They note that the user surplus from the US-China trade relationship is still significant… They also acknowledge that this trade relationship may be disrupted, potentially hurting companies like Apple. They suggest that moving China's low-end manufacturing to places like India and Vietnam may be a good thing for the world. They also discuss the possibility of restricting exports of semiconductors to China, but note that this may not be a net positive. They then shift to discussing “key insights”, specifically the concept of “Hexapodia”.. They then discuss the “polycrisis”, a term coined by Adam Tooze, which refers to the interconnected issues of friend-shoring, decoupling, deglobalization, and other crises.

Finally, they briefly touch on pronunciation and the challenges of correctly pronouncing foreign words and names.

They [move on to] briefly discuss Hawaiian independence, noting that while it was long ago, there is no desire for it among the current population…. They suggest that territories should either be independent or become states, with no in-between designation. They then shift to the topic[s] of friend-shoring, decoupling, and declobalization… a big ball of wax with significant implications for the global economic system. They note that the long-standing belief that lowering tariffs and increasing trade would reduce the chances of war has been successful, but there are still challenges to be addressed in the current global economy.

Noah Smith and Brad DeLong [next] focus is on the idea of trade and the relationship between the United States and China. They discuss how people become friends with those they trade with, but only if they think they are trading fairly. The Chinese Communist Party believes that the trade deals struck with the US gave them too little of the surplus, while the US believes it gave away too many blue-collar jobs. The discussion then turns to a recent incident involving a Chinese spy balloon, which many Americans are claiming is a big deal, while others are puzzled by the reaction. The hosts suggest that Americans often get worked up over minor things while ignoring more important issues. They also note that China has become a toxic thing in American culture, and that many people view China as an enemy. The hosts wonder if this sentiment is here to stay for a while and note that the US did not do a lot of trade with Russia when they were considered an enemy….

DeLong [then] expresses gratitude for his laptop and iPhone, which were assembled by people in China using components from Taiwan and Korea. He compares this to a gift-exchange, where he feels he has received more than he has given. However, DeLong notes that the traditional morality of gift exchange does not include theft, which is what China has been accused of with regards to intellectual property theft from the US. He acknowledges that there may be arguments for a poorer country to steal to catch up technologically, but it still constitutes theft. Smith notes that the US engaged in similar theft of British intellectual property in the 1800s, due to not liking British intellectual property laws. They also touch on the US-UK relationship, which started as a rivalry but quickly turned into a close friendship during World War I and II, with the US saving Britain's bacon both times. They also mention that China was among those whose bacon was saved during World War II.

Noah Smith and Brad DeLong [then] discuss how the US has helped China in the past, but China seems to have forgotten this assistance… during World War II and again in 1969 when the Soviet Union threatened to launch a nuclear strike on China [and] the US made an alliance with Maoist China… which helped hasten the end of the Cold War and led to the transfer of military technology to China. DeLong acknowledges that the US had a negative impression of China after the Korean War, but tried to shepherd China into becoming a responsible new superpower. However, China now seems to view the US as a bad enemy that it should be planning to defeat and overthrow. DeLong believes that China would not be where it is now without assistance from the US, and that this assistance will be an important asset in solving global issues in the future. He argues that virtue is often rewarded, even if it isn't, and that the memory of the US's assistance to China will be important in achieving a free and democratic society on a global scale.

Noah Smith and Brad DeLong [next]… discuss the failure of the engagement approach, and the possibility of China escaping from the middle income trap. Brad DeLong suggests that a more comfortable world for the US to live in is one where China does not escape from the middle income trap… [but that] would be worse for the world. They also talk about the possibility of China becoming a technological superpower and how this will depend on what happens over the next two generations. They conclude by discussing the pronunciation of various words in different languages.

They both agree that China has structural incentives against becoming a truly rich and prosperous country, including its lack of political and economic freedoms. They also discuss the potential for de-globalization…. They suggest that a future de-globalization could occur if there are rapid advances in robotics, leading to more manufacturing jobs in the US and other countries…. They note that trade as a percentage of GDP has been decreasing slowly since before the financial crisis, and while de-globalization is happening, it is not happening rapidly. They also discuss the idea of "friend-shoring," which is not the same as de-globalization, but rather a shift in globalization.

They caution against taking individual human relationships as an analogy for relationships between countries, as countries have no gratitude or memory. They express pessimism about the current era of tension between the United States and China and the potential for it to escalate into a new Cold War…. It may take different leadership in both countries and time for the bitter social divisions in the United States to heal before relations between the two countries can improve.

Overall, the tone is pessimistic, but they end with a lighthearted discussion of hexapodia and the dangers of ferns on six-wheeled scooters.

I judge this as considerably more useful than the voice-recognition transcript of the audio, at least given the current state of voice-recognition technology.

Does it get me materially closer to actually having a memory—to being able to surface the relevant and thoughtful things from the stock of what I have written and said in the past, and do so when it is useful to me and my concerns (or to others and their concerns) in the present?

Perhaps.

I can say that this makes me materially more optimistic about the prospects for finding this particular Holy Grail before I die…

But what I simply do not understand is what keeps Chat-GPT3 from hallucinating here. By contrast, its hallucinations are remarkable when I ask it: What are the ten most-recent writings economist Brad DeLong is best known for?

Brad DeLong, an economist and professor at the University of California, Berkeley, has authored numerous publications on a wide range of economic topics. Here are ten recent writings he is known for:

1. "Productivity and Potential Output Before, During, and After the Great Recession" (2021): This paper analyzes the impact of the Great Recession on productivity and potential output in the United States.

2. "The Slump and the Aftermath" (2020): This paper examines the causes and consequences of the 2008 financial crisis and the ensuing Great Recession.

3. "Long-Term Damage from the Great Recession in OECD Countries" (2019): This paper assesses the long-term effects of the Great Recession on labor productivity and potential output in advanced economies.

4. "Should-Read: Understanding Income Inequality in the United States" (2019): This blog post discusses the causes and consequences of income inequality in the United States.

5. "The Productivity Puzzle and Misallocation: An Italian Perspective" (2019): This paper explores the causes of the productivity slowdown in Italy and the role of misallocation in the Italian economy.

6. "Macroeconomic Policy in a World of Second Best" (2018): This paper argues that policymakers should aim for second-best solutions in macroeconomic policy, rather than pursuing unattainable first-best outcomes.

7. "The Political Economy of Globalization in the United States" (2018): This paper examines the political and economic factors driving globalization in the United States and its impact on income distribution.

8. "The Future of Work in the Age of Automation" (2017): This paper discusses the potential impact of automation on the future of work and employment.

9. "The Long-Run Fiscal Trajectory of the US: From Optimism to Crisis to...?" (2017): This paper analyzes the long-run fiscal sustainability of the United States, given projected demographic and economic trends.

10. "The Scary Debate Over Secular Stagnation" (2015): This blog post discusses the concept of secular stagnation and the policy implications of the idea that the economy may be stuck in a period of low growth and low interest rates.

All hallucinations except for (10)—and ten is not a blog post, but an article in the Milken Institute Review.

---

ONE IMAGE: Experience Machines:

You—one—or, at least, I—do not need VR!

Very Briefly Noted:

• Sidney Pollard (1981): Peaceful conquest: the industrialization of Europe, 1760-1970… 

• Jesse Rothstein & Till von Wachter: Social Experiments in the Labor Market: ‘Many important questions… require going beyond random assignment… endogenously observed outcomes, such as wages and hours; spillover effects; site effects; heterogeneity in treatment effects; multiple and hidden treatments; and the mechanisms producing treatment effects…

• Enda Curran: What If We’re Aiming for the Wrong Inflation Goals?: ‘“If people sat down today they would not come up with 2%, they would come up with 3% to 4%,” Mohamed El-Erian, the chairman of Gramercy Funds and a Bloomberg Opinion columnist told Bloomberg Television…

• Marcy Wheeler: BREAKING from Fox News: Trump Cheated ... and He Still Couldn't Beat Joe Biden: ‘Rupert Murdoch, in a sworn deposition as part of Dominion’s lawsuit against Fox, confessed that he helped Donald Trump cheat during the 2020 election. He provided Jared Kushner confidential information about Joe Biden’s ads, the kind of information that Trump had to rely on Russian spies to obtain from Hillary in 2016…

• Dominion Voting Systems: Combined Opposition to Rule 56 Summary Judgment Motions RSM 

• Zoltan Barany: Armies and Autocrats: Why Putin’s Military Failed: ‘Putin’s monopolization of control over the armed forces, which has driven critical voices and honest debates out of military and defense matters… the failure of reform… unable to attract talented young people…. The Russian military is a reflection of the state that created it...

• Michael Ignatieff: The Politics of Enemies: ‘Even if Americans agree about the procedural content of democracy, some apparently no longer trust their officials to carry out these procedures…. What was unique… [was] the willingness of some elected representatives… to side with insurgents and later to excuse the desecration…

• Adam Garfinkle: The Age of Spectacle: ‘US political culture may now be described as an addictive spectocracy...

¶s:

Matt Levine: ChatGPT: ‘The stereotype of chat-based large-language-model artificial intelligence products in early 2023 is that they are very good at sounding smart but not so good at being correct…. Given that stereotype, you might imagine two ways for big banks to think about these models: “Being fluent, confident and wrong is a core job function of a banker or investment analyst, and ChatGPT allows our employees to be much more efficient at it, so we should encourage them to use ChatGPT as much as possible.” “Our clients will be annoyed with us if we are constantly fluent, confident and wrong, so we should ban use of ChatGPT until we are more confident that its recommendations are correct.” Obviously the first approach would be very funny, but the second is more realistic. But in fact there is a third, less obvious, but even more correct way for big banks to think about the use of these models: “If our employees type about business on a computer or a phone, it had better be in a software system that we control and that creates a searchable record that we can preserve forever, because otherwise our regulators will get mad at us. ChatGPT is an artificial intelligence model, but it is also a box for typing on a computer, and that’s too big a regulatory risk for us.”… If you want to get advice from a robot about how to invest — or if you want the robot to help you write a presentation for clients — then you had better communicate with the robot using official channels! Typing in the ChatGPT box isn’t an official channel, so it's not allowed…

---

Tom Nichols: The war in Ukraine is the end of a world: ‘Personal grief about the passing of the hopes so many of us had for a better world at the end of the 20th century…. I was wrong. I underestimated the power of Soviet imperial nostalgia. And so today, I grieve. I grieve for the innocent people of Ukraine, for the dead and for the survivors, for the mutilated men and women, for the orphans and the kidnapped children. I grieve for the elderly who have had to live through the brutality of the Nazis and the Soviets and, now, the Russians. I grieve for a nation whose history will be forever changed by Putin’s crimes against humanity. And yes, I grieve, too, for the Russians…. Finally, I grieve for the end of a world I knew for most of my adult life… [a] world… full of chaos, but… also grounded in hope…. After 1991, time seemed to be on the side of peace and democracy, if only we could summon the will and find the leadership…. Now I live in a new era…. Democracy is under attack everywhere...