I Made Five Bucks

AI on the Current Conflict

My Recent Podcast

My recent podcast guest on Brave New World was with Rajesh Jain, one of India’s original Internet entrepreneurs. Rajesh made out like a bandit in the 90s after selling his Internet platform company for over a hundred million bucks, and started thinking bigger. He used his platform to help India’s current prime minister, Narendra Modi, in his successful election campaign in 2014. Since then, he’s tried to use his digital outreach capability in creative ways, like trying to improve India’s democratic process by helping promote dialog and surfacing candidates who mightn’t otherwise become visible. That hasn’t quite worked out as intended, but perhaps it’s a question of timing and perseverance, and not entirely a quixotic endeavor. He’s also exploring the use of his digital targeting platform to get  data-based snapshots of things like public mood on important issues.  Andrew Yang drew attention to the need for such types of indicators in my conversation with him last year.

What I find interesting about people who’ve enjoyed huge success is the many failures they’ve endured and learned from along the way. It helps to have a Pepe La Pew attitude: if you fail, try again, and again, and again. But unlike Pepe, learn.

So, check out my conversation with Rajesh!

Pain to Gain

I always get a chuckle recounting a story of a friend and former New York Mercantile Exchange floor trader, who went from euphoria to near destruction within minutes. He survived, ending the day flat, and stumbled home in a daze. His wife asked him about his day.

I made five bucks, he said.

Oh, quiet day, she said.

There’s an old adage that goes no gain, no gain. But it’s a shame when there’s pain without any gain.

Gain and pain are central concepts in life. They’re key in investing, where they are objectively measurable. For example, say your strategy is to invest in the strongest American companies, represented in the S&P500 index. You can calculate the gain over a period, by say, considering the average daily return of the index over that period. To measure pain, consider what my friend went through: a dizzying and volatile ride. Volatility is mathematically measurable, for example, as the dispersion of returns around the average. The wider the dispersion, the more dizzying the ride.

I call the ratio of the two the “Gain to Pain” ratio.

The gain to pain ratio is the standard way to measure risk-adjusted performance in the industry, where it is referred to by a fancier name: the “Information Ratio.” It usually has a nice and simple interpretation: a ratio of 1 implies you have a one in six chance of having a losing year. A ratio of 2 reduces that probability to roughly one in twenty. A ratio of 0.5 is roughly one in three.

As an example, Warren Buffett has had a gain to pain ratio of 0.56 over the last five years, marginally higher than the S&P500’s ratio of 0.53. Microsoft and  NVIDIA have had a ratio of roughly 1. You can measure yours easily by using your monthly brokerage statements. It’s an interesting exercise, and might change how you think about gain and pain. You might realize that past gain doesn’t usually predict future gain, but past pain often predicts future pain!

I’ve found gain and pain to be a good way to think about a variety of situations that involve calculating the “average goodness of a stream of things.” I used the thinking recently in a research project for a large insurer to calculate the relative goodness of their various lines of business. Interestingly, their gains were more predictable than their losses, but after adjusting the math accordingly, it was an interesting exercise to be able to compare the relative goodness of their various lines of business and calculate the probability of a losing year in each.  It provided  the chief with a new way of thinking, and I learned about yet another application.

I’ve found gain and pain to be a great way to compare alternative choices in general, and I re-use the thinking all the time, as in the case above. It reminded me about my conversation with Chandrika Tandon a few weeks ago who cut her teeth as a young McKinsey consultant on Banco de Brazil, and reused the learning on Chase and other institutions. It’s a critical success factor in consulting: learn and reuse. I find myself thinking about how to estimate gain across all kinds of problems.

Gain to Pain in Conflict

The precise estimation of pain and gain is impossible in social situations and those involving conflict, such as the war between Israel and the extremist Palestinians. The emotional pain of losing one’s home or a loved one is impossible to quantify, as are the gains from killing people.

There’s an old Golda Meir quote: “we have no place to go.”

The Palestinians have nowhere to go either. It’s not like they will disappear somehow into the Arab world, and the problem will go away. Many were displaced from land they previously occupied.

Game Theory offers a framework for addressing conflict when the outcome of one’s choice depends on the response of the other party. Opposing choices must be estimated. The goal is to find an “equilibrium” situation, where the actions of both parties lead to an outcome that cannot be improved upon by other choices.

In the Israel-Palestine conflict, an equilibrium outcome is for both parties to share the land. They have nowhere else to go, so the alternative is annihilation or perpetual conflict. Unfortunately, Game Theory breaks down when too many parties are involved. In this case, it is complicated by other parties such as Hamas, who will disrupt any move towards equilibrium.

In a recent podcast with my colleague Scott Galloway, the political commentator Fareed Zakaria made two salient observations about the current conflict. First, this time around, Hamas is literally willing to burn the house down. It’s an act of desperation, where they anticipate severe repercussions as part of their pain calculation. Zakaria’s second point was that in a conflict where one party is much stronger, the negotiating position of the weaker party only gets worse over time. That’s what’s been happening to the Palestinians. Relative to the almost successful President Clinton-brokered agreement between Yassar Arafat and Ehud Barak over 20 years ago, Zakaria noted the Palestinians are down to “40% of the original loaf of bread” they could have gotten in 2000, and the loaf keeps getting smaller over time as Israeli settlers continue their expansion.

Zakaria puts the blame on the failure to find an equilibrium solution largely on the Palestinian leadership, who failed to grasp the fact that when you’re dealing with a superior enemy, stalling only reduces what the other party will be willing to offer you in the future. Eventually, there will be no gain, only pain. That’s what they’re up against. And without a wiser leadership, their prospects will only wither.

What Does the AI Say?

I decided to ask GPT what it would recommend as a solution to the conflict, to see whether the machine’s lack of emotion would offer a fresh perspective. Recognizing that it is only trained on data until 2021, I began by describing the current context as best I could as a hypothetical situation, and asked it for advice.

GPT said Israel should condemn the terror attack, and address the root cause of the conflict, including historical grievances and land disputes, be patient, and willing to compromise. It said that the Palestinians should seek international support to stop the bombing of civilians, and garner sympathy from the international community to recognize and support their cause. It encouraged both sides to pursue diplomacy. Nothing surprising there.

I told it that such attempts had failed over the last few decades, so what should we do now?

Here are its suggestions for both sides:

·       Find a potentially better neutral party than those in the past, who may not have been perceived as being sufficiently neutral.

·       Engage in prisoner-exchanges and other measures for building goodwill.

·       Encourage unofficial, "Track II" diplomacy involving civil society organizations, academics, and individuals from both sides to work on building bridges and finding common ground.

·       Support initiatives that promote education and reconciliation, particularly among younger generations. Changing attitudes and building trust take time and often starts with the youth.

·       Mobilize public opinion and international pressure to encourage the parties to seek peaceful solutions and make concessions.

·       Reexamine and address the underlying historical grievances and land disputes with a fresh perspective, seeking creative and mutually acceptable solutions.

·       Support efforts to rehabilitate and de-radicalize individuals who have been involved in extremist activities. Rehabilitation can be a critical component of preventing further acts of terrorism.

I thought that the AI’s suggestions were constructive. It suggests broadening the dialog beyond traditional back-channel diplomacy to include academics and large numbers of young people, including students. Perhaps the leaders currently involved, especially Palestinian leaders, are failing to represent their people who want space to live and a better life.

It is ironic that in one of the thorniest issues of the day, AI seems to be displaying more wisdom than many humans whose emotions are running high. Its implicit gain to pain calculus suggests that the path forward for both sides is negotiation, not escalation.