Swaine’s World
Jerry Kaplan: the Need to Know Interview
Featuring Tech Tales, Poetry Corner, and More
Like Swaine’s World? Scroll to the bottom of this post to tip the author or to subscribe.
About the Picture
The pictures at the top of these posts rarely have anything to do with the content of the post. Since this blog is Swaine’s World, I try to show some aspects of my world, mostly scenes around my home. This particular one is to remind me, as the temperature hits triple digits four days straight, that we also have days like this down at the river.
Jerry Kaplan: the Need to Know Interview
In 2016, Jerry Kaplan published Artificial Intelligence: What Everyone Needs to Know. Now he has written another What Everyone Needs to Know book, dealing with generative AI, due out this fall from Oxford University Press. Kaplan has a unique perspective on AI, from bringing AI capabilities to commercial software forty years ago to lecturing these days at Stanford on modern AI. Paul Freiberger and I wanted to get his long-view insights on modern AI for the next edition of Fire in the Valley. Here are highlights of that interview, prefaced by some background on Kaplan and the earlier book.
Jerry Kaplan: What Everyone Needs to Know
Some details from Wikipedia:
“Kaplan is widely known as a serial entrepreneur, technical innovator, bestselling author, and futurist. He co-founded four Silicon Valley startups, two of which became publicly traded companies. His best-selling non-fiction novel Startup: A Silicon Valley Adventure was selected by Business Week as one of the top ten business books of the year.
“Kaplan is currently a Fellow at The Stanford Center for Legal Informatics. He also teaches ‘Philosophy, Ethics, and Impact of Artificial Intelligence’ in the Computer Science Department, Stanford University. He holds a BA (1972) from the University of Chicago in History and Philosophy of Science, and an MSE (1975) and PhD (1979) in Computer and Information Science, specializing in Artificial Intelligence and Computational Linguistics, from the University of Pennsylvania.
“Products co-invented: The Synergy (first all-digital keyboard instrument, used for the soundtrack of the movie TRON); Lotus Agenda (first personal Information manager); PenPoint (tablet operating system used in the first smartphone, AT&T’s EO 440); the GO computer (first tablet computer) and Straight Talk (Symantec Corporation’s first natural language query system). He is also co-inventor of the online auction (patents now owned by eBay) and is named on 12 U.S. patents.
“Published research: He has published papers in refereed journals including Artificial Intelligence, Communications of the ACM, Computer Music Journal, The American Journal of Computational Linguistics, and ACM Transactions on Database Systems.”
From his earlier book on AI:
“Over the coming decades, Artificial Intelligence will profoundly impact the way we live, work, wage war, play, seek a mate, educate our young, and care for our elderly. It is likely to greatly increase our aggregate wealth, but it will also upend our labor markets, reshuffle our social order, and strain our private and public institutions. Eventually it may alter how we see our place in the universe, as machines pursue goals independent of their creators and outperform us in domains previously believed to be the sole dominion of humans. Whether we regard them as conscious or unwitting, revere them as a new form of life or dismiss them as mere clever appliances, is beside the point. They are likely to play an increasingly critical and intimate role in many aspects of our lives.
“The emergence of systems capable of independent reasoning and action raises serious questions about just whose interests they are permitted to serve, and what limits our society should place on their creation and use. Deep ethical questions that have bedeviled philosophers for ages will suddenly arrive on the steps of our courthouses.”
The future he predicted has arrived. Here are some highlights of our interview, lightly edited for better bloggification.
How It All Changed Overnight
“It was really around last November, less than a year ago, that these systems just blossomed. AI has been plodding along for years, making significant progress on certain classes of problems. Things had been getting better and there were new techniques and all that. And then about five years ago, a new technique was developed, a new architecture for these systems, called a transformer.
“It’s not that big a departure of what was from before. It just gave [the system] a way to respond to questions much more efficiently by directing its attention to those parts of a conversation that are particularly relevant to understanding what to say next.
“And several other advances converged. There had been advances in natural language processing, which [had gotten] pretty good. They were working on translation and other issues. But a new way of representing meaning arose about five years ago in a concept called embeddings.
“The way that it represents the meaning of words in these systems turns out to be an incredibly powerful innovation. It wasn’t even recognized by the people doing it. It was just an interesting idea that allowed them to solve certain kinds of language problems.
“So [the first thing] is the invention of word embeddings. The second thing is the change in architecture to a transformer architecture. The third is just the inevitable march ahead in computing power. And once these things started to look good, if you threw enough money at it and enough processing power, they really improved quickly and had these so-called emergent properties.
“And now we’ve got something that’s really new and different, and it’s going to have an enormous impact on society. I mean, this is going to, in my view, certainly equal and probably exceed the impact of the personal computer. And this is generative artificial intelligence.”
Beyond Chatbots
[We cited a recent article in the New York Times about work at Google where they’re tying these large language models to vision, technology, and robotics. We asked, doesn’t this make generative AI more scary than just chat bot stuff?]
“Yeah. No, see, people don’t get this. The amazing stuff that we’re seeing today is just an appetizer for a big meal. This is what happened to these devices when they just scraped up all the crap off the internet and fed it in: just based on words, it can now solve all kinds of problems and engage in all kinds of intelligent behavior and creative behavior nobody, I think, ever really expected.
“And what you said is true. This is just based on the words. When we hook it up to real sensors in the real world and allow it to take actions directly rather than just spitting back written answers to things, there’s no telling where this is going to go.”
What Have We Created?
[We asked about the philosophical meaning of generative AI. Have we engineered the mechanism by which the human mind works? Have we created a monster, a new form of life?]
“There are certainly some characteristics that have emerged of these systems which have parallels in human behavior, but this is not an artificial brain, and the systems have limitations that I think really suggest that this is something that’s completely new and different.
“So no, I don’t think this is some kind of new form of life. No, none of that. There’s no evidence for any of that. This is a tool. It’s a computer program and it does stuff. It just does it incredibly well. And it probably does shed some light — over the next 10 years or so — as to what are we and how do we function and what does it mean to be human.
“The way I’ve been describing this is this is a tool that is smart enough to use tools. So predicting what it can do is very difficult, because it can use all the tools that previously only people had any idea how to use. And more than that, it’s an invention that can invent.”
[So if generative AI is a tool, then we need to focus on how this tool will be used, right?]
“in my new book, I’ve got several chapters on just this question, because now it’s really going to be an issue. The issue of how these new AI systems that we’re building should… The term is alignment. They call it the alignment problem. Whose interest should they serve and how should they moderate their behavior to accommodate normal human conventions for cooperation and sharing, and not just ruthlessly pursue the goals of one individual, one person. It’s a very interesting question and a very difficult technical problem.”
Indeed.
Next week: a visit to the past as I unearth an old “Swaine’s Flames” column from my Dr. Dobb’s Journal days. But now, this:
Tech Tales
The Tech Tales series is a history of the computer, lightened with semi-fictional dialogs and delivered in bite-sized chunks over the course of 2023. Here’s this week’s offering.
Scene:
Bob Marsh and Lee Felsenstein are drinking beers at the Oasis, the favorite watering hole of the members of the Homebrew Computer Club. Although Bob is now president of a computer company raking in millions of dollars and Lee is the designer of the company’s computer, they still hang out and share technical insights with their friends and sometimes competitors at the club meetings. Tonight, Bob is worried.
Bob: They’re coming, you know.
Lee: [Looks around the crowded bar] It looks like they’re already here.
Bob: I mean the Big Boys. They’re going to jump into the market one of these days.
Lee: Ah, right. Attracted by the smell of money.
Bob: It’s been fun, brainstorming with the other Homebrewers and building a company and cashing the checks. But it’s all going to change. And soon.
Lee: No question. The question is, who will survive after IBM or Commodore or National Semiconductor comes out with their personal computer and crushes the little hobby companies.
Bob: That’s why we have to get really big really fast. The gate is coming down, and we need to be on the other side when it does.
Lee: Remember what Texas Instrument did to Ed Roberts in the calculator market.
Bob: Who do you think it will be? Which big company is going to move in on us first?
Lee: Could be any of them. Oh, god. Anybody but TI!
The Hobbyists and Their Greatest Fear
A group picture of the executives of the personal computer industry at the end of the 1970s would show mostly young male electronics hobbyists. Some would be wearing suits, with wide lapels and wider ties, some would be in short-sleeved shirts with multiple pens in pocket protectors, some in t-shirts and jeans, possibly one or more barefoot. It was still an industry of self-styled hackers, but they were starting to wonder how long that would be the case. They could almost see the gate closing. It was becoming way too obvious that there was money to be made here. Apple was seeing to that.
Steve Jobs, as the public face of Apple, would be out in front of this imaginary group picture. Apple was firmly rooted in this hobbyist community, but it was playing the game like a big company. There was the quarter million dollar Markkula investment and his aggressive business plan. Markkula seriously intended to turn Apple into a Fortune 500 company in five years, and if anybody could do it, he could. Then there were the ads, big, attention-grabbing ads in attention-grabbing places, like Playboy magazine.
Apple hardly needed to advertise: the press was happy to write about them, usually in glowing terms. Their computer, the Apple ][, looked nothing like the metal boxes everybody else was selling: they had sprung for a sleek, beige plastic case, designed by an actual industrial designer. Apple in every way conveyed that personal computers, theirs at least, were serious consumer products. Home appliances, not toys for hobbyists. And Apple was selling enough Apple ][s to prove that this niche market of “personal computers” was the real thing. In fact Apple was creating a nice market for add-on products. Software, especially, was a growing market. VisiCalc was the breakthrough product, proving that personal computers could be useful for business.
But to Apple and all the other players, it felt like only matter of time before the big boys came in and spoiled it all. Ed Roberts was not the only one who remembered how Texas Instruments, or TI, had stepped into the calculator market a few years earlier, cut prices to the bone, and destroyed the business of TI’s own customers. It could happen again. Why wouldn’t it? Intel, it’s true, had made it clear from the start that they opposed to competing with their own computer company customers. But TI had demonstrated that it had no such qualms, and there were already rumors of a possible TI computer. National Semiconductor was another semiconductor company supposedly working on a computer. When they stepped it, it would be all over.
It didn’t even have to be a semiconductor company. There were other big boys. Like the mainframe and minicomputer companies, who you just knew had to be looking at this personal computer market and thinking, this is our rightful domain, even if it is currently occupied by these scruffy squatters. Or the big electronics companies, with their established distribution channels or their own stores. They weren’t just rumors, they were existing threats. Commodore International had released a personal computer called the Pet that was popular in schools. Radio Shack was also already in the market. They were a retailer a network of stores with no track record in manufacturing. Nevertheless they had launched a personal computer and were competing well in the market, with sales comparable to Apple’s.
The Big Boys were coming.
To be continued.
The Poetry Corner
Shell Game
Lottie played
Luigi’s game,
Paid in queer
And split for Maine.
Brady was
Luigi’s shill,
Took his pay
In funny bills.
When he saw
The bread was phony,
Sent it off
For alimony
To his old lady,
Lottie Brady.
From my first collection of verses, cleverly titled First Verses. You can download First Verses for free.
Tune in next week for more adventures in tech history and light verse.
