Isegoria

Increasingly, Ethan Mollick notes, work at OpenAI looks like managing AI:

A quarter of OpenAI workers have at least four agents running at one time every week. And, as coding is done by AIs in specialized harnesses and apps, other roles start to become coders of a sort. And they are good at it.

A separate study of Claude Code users found that software engineers had a similar success rate to other professions when actually using Claude code on coding tasks.

What actually mattered was not the profession of the user, but their expertise. The more domain experience someone had, the more successful they were in using Claude Code in that domain. And, even more interestingly, the more useful output they got from Claude from each prompt.

We are moving from a world where non-experts use chatbots to fill in gaps to one in which experts use agents to get work done. And the best way to use agents is to think of yourself as a manager.

In the spring of 2024, David Epstein explains (in Inside the Box), he met up with a group of General Magic alumni to discuss their proto-iPhone:

It had been three decades since they went public—with no revenue—in Silicon Valley’s first “concept IPO.”

[…]

On the day of the IPO, Sony’s investment in General Magic instantly appreciated so much that it was more than they had anticipated making in their five-year business plan for the product. “And then it was gone,” Gruzen said. “For a moment, it was fantastic.”

[…]

Professor Tom Eisenmann, author of Why Startups Fail, rolled in a whiteboard so that the General Magic group and the audience could compile a list of what had gone wrong. The board filled up quickly. Too many partners; too early; too little attention to the limits of available technology; too vague a customer; too much making everything on their own instead of building on others; too little management. One audience member, an entrepreneur and mentor to Harvard and MIT students, described at length how watching the film reminded her of her mentees, whose giant visions often get in the way of accomplishing anything at all. Professor Eisenmann distilled her wisdom on the whiteboard: “Do something, not everything.” Laughter and head nods spread around the room.

Steve Jarrett explained to the audience that the software engineers were so skilled that they could imagine and build new things literally overnight. “We just kept inventing new ideas, and it kept making the product bigger, and bigger, and bigger.” The CEO, he said, “would just tell a new story every time there was a new idea.”

Jarrett recounted how one of the third-party app developers he managed, Jeff Hawkins, did what they could not—something, not everything. Hawkins had designed an app called Graffiti that could be used on General Magic’s platform. It allowed users to make quick strokes with a stylus that would be recognized as letters and punctuation. “So Jeff called and said, ‘How many devices have you sold?’ ” Jarrett explained. Jarrett shared the disappointing news, to which he recalled Hawkins replying: “Oh my God.” After that, Hawkins and his small team decided to build their own, simpler product. “When he told us that, we said there’s no way. How are you going to compete with Sony and Motorola?” Jarrett recalled. “And he crushed us, because he took this simple iterative approach to solving problems.”

Hawkins took his one tiny component of General Magic’s giant ecosystem and made it the basis of a much simpler device. It started just with functions for a calendar, contacts, a to-do list, a memo pad, and a calculator. That was the PalmPilot, which became a smash hit in the same era.

[…]

“We had too much on our plate,” Stern told me. “If you don’t whittle it down, you’re not gonna make it. And we kept piling it on. Why did we need sixteen partners? We didn’t. But there was all this FOMO.”

In Why Startups Fail, Eisenmann describes six patterns that account for the vast majority of startup failures:

Bad Bedfellows. Startup success is thought to rest largely on the founder’s talents and instincts. But the wrong team, investors, or partners can sink a venture just as quickly.

False Starts. In following the oft-cited advice to “fail fast” and to “launch before you’re ready,” founders risk wasting time and capital on the wrong solutions.

False Promises. Success with early adopters can be misleading and give founders unwarranted confidence to expand.

Speed Traps. Despite the pressure to “get big fast,” hypergrowth can spell disaster for even the most promising ventures.

Help Wanted. Rapidly scaling startups need lots of capital and talent, but they can make mistakes that leave them suddenly in short supply of both.

Cascading Miracles. Silicon Valley exhorts entrepreneurs to dream big. But the bigger the vision, the more things that can go wrong.

General Magic was already in trouble, David Epstein explains (in Inside the Box), when a tech-support guy named Pierre Omidyar used his free time and personal website to start AuctionWeb:

General Magic had a right to any IP developed by employees, so Omidyar dutifully showed the company’s lawyer that his site was generating traffic, connecting people, and facilitating commerce. Isn’t that what General Magic wanted to do? But they were already too committed to the partners of the gigantic Alliance and to a proprietary e-commerce network to change direction, even though the explosion of the internet should have completely altered their plans. General Magic passed on AuctionWeb. Omidyar, who didn’t have a business bank account and had been collecting checks in a pile at his desk, kept it going. He left General Magic, and changed the name of his site to eBay.

Larry Tesler, who was leading Apple’s Newton group, hired an anthropologist named Eleanor Wynn, David Epstein explains (in Inside the Box), to test the theory behind their business plan:

Wynn reported back that business­people weren’t going to use the Newton unless it included a phone. But adding a phone at that point (as General Magic was busy learning the hard way) was unreasonable. In that case, Wynn reported, the market will be workers who already carry separate pieces of communication equipment, like police officers and firefighters. In an oral history interview for the Computer History Museum, Tesler recalled Apple CEO John Sculley’s response:

Sculley is deflated, “I thought my market was CEOs, like me. Not firemen.” And “No, no, no, we don’t want that. That answer is not acceptable.”

So they went hunting for another answer, using what Tesler called the “famous fake focus group.” Tesler was told that the supposed focus group would be in Minnesota, because people in the Bay Area were too tech savvy, and neither he nor Wynn were invited, and unfortunately there would be no recording of it they could watch. The focus-group team came back with a message: “We’re building the exact right product.” Apple plowed ahead, and the Newton flopped spectacularly.

Bell Labs is often remembered as the epitome of unfettered exploration, David Epstein explains (in Inside the Box), but that framing misses a crucial point:

Eric Gilliam, who studies and writes about innovation history, coined the beautiful phrase “long leash, narrow fence” to describe the ethos at Bell Labs in its heyday. New researchers were given extraordinary latitude in determining what to work on, but were expected to interact with engineers and manufacturing facilities to identify specific problems that needed solving.

[…]

John R. Pierce, a Bell Labs scientist and “father of the communications satellite,” recalled in an oral history: “Too much freedom is horrible. It’s like telling a young child, ‘Do whatever you want to.’… It’s certainly bad to be directed to do things very, very narrowly and with no freedom. It’s my guess that for every person who needs more freedom, there are ten people who need more help in finding their way.” What they had at Bell Labs, as another famous scientist put it, was “circumscribed freedom”—freedom within a framework.

Gilliam shares just an excerpt from an interview with Pierce, from 1979, after he had earned acclaim as the father of the communications satellite:

LYLE: I want to talk about research in the Bell Labs and how that’s done. That is, when you first started there, you were working with vacuum tubes. Who decides what problems will be worked on?

PIERCE: That’s very different then and now. I was told to do research on vacuum tubes. People sort of just left me alone. They did suggest that I go and see Philo Farnsworth, who was working on electron multipliers and television pick-up tubes, but I was left pretty much to myself. This was very, very confusing to me. I didn’t know what to do.

LYLE: Were you doing it alone?

PIERCE: Yes.

LYLE: Did they say, “So-and-so has been doing this and this is where he left off”?

PIERCE: No. I was just supposed to plan something to do and do it. I think that is close to cruel and unusual punishment.

LYLE: And full of anxiety, I’m sure.

PIERCE: Yes, but I didn’t know enough to be unhappy. I did crazy things. I did some useful things. I invented an electron multiplier. I was greatly helped at this point, but not so much by the people who were close to such work. I felt a certain secretiveness in the people who were working near to me. They were doing their own thing, and I was doing other things. Heaven knows how I found anything useful to do. I was exposed to things by some of the people who were less secretive. I was very much helped by Bill Shockley, who came to Bell Laboratories about the same time I did. He had been an undergraduate at Caltech but did his graduate work at MIT. He was a very sympathetic person, and taught me a good deal. Somehow I hit on things that were worth working on — electron multipliers and the question of noise in electron multipliers, and later trying to make high transconductance vacuum tubes.

Then, as the war came, I was drawn into microwave tube work, and the outcome of that was a little bit by accident. First, I tried to make klystron amplifiers—I’d heard about klystron. Then I stumbled onto reflex klystrons, which was not a new idea, but I stumbled onto it independently. Gerry [William Gerald] Shepherd, who’s now at the University of Minnesota, and I made some klystrons that were in all American microwave radar receivers. The magnetron was the big thing of the day, but we made these beating oscillators for receivers instead.

Too much freedom is horrible. It’s like telling a young child, “Do whatever you want to.” You’ve heard this story. There are various outcomes. One is, “Do I have to do what I want to?” Complete freedom is not very helpful to a person who is inexperienced in the world. It’s certainly bad to be directed to do things very, very narrowly and with no freedom. It’s my guess that for every person who needs more freedom, there are ten people who need more help in finding their way.

LYLE: So, did they tell you why they wanted the vacuum tubes, when you started off?

PIERCE: Not really. I found out some way, inadvertently. Some people were working on electron multipliers, and I made some improvements on them. It became clear that people needed better vacuum tubes for building negative feedback amplifiers, and I worked on that. I don’t think I was told this formally; I just found out by talking to people. Then, as the war approached and we got into war, it became apparent that microwave radar was very, very important, and I worked on tubes for radar. It was a process of osmosis rather than direction that led me into these things, as I remember it.

LYLE: How was the research tied in with the general business of Bell Telephone? kind of a relationship exists between these two parts of the company?
That is, what kind of a relationship exists between these two parts of the company?

PIERCE: It’s a very important relationship. The Bell System has AT&T, which is sort of a holding company, but it also runs the long lines that provide long distance telephone service. It establishes engineering practices for the Bell System. It owns Western Electric, which is a manufacturing organization, and it also owns, together with Western Electric, the Bell Telephone Laboratories.

I remember that during the war we saw a good deal of people from Western Electric, who were going to manufacture the things that we devised. Because all of these people were engaged in telephony, or during the war because they were all engaged in radar and other military things, you got to talk to people who were engaged in the operation of things, who were engaged in the manufacture of things, and you got a picture of the rest of the world which certainly influenced what research you did.

I can understand a university, which does teaching and research. But the idea of a research institute without ties to either teaching or to manufacturing or operational organization seems a terribly sterile idea. You see that in the Soviet Union; there’s a lot of good activity that never results in anything. When they want to build automobiles, they hire Fiat to build an automobile plant, instead of relying on what they have learned.