Few have thought through the costs that will impose on those who are supposed to respond to all this new text. One of my favorite examples of this comes from The Economist, which imagined NIMBYs — but really, pick your interest group — using GPT-4 to rapidly produce a 1,000-page complaint opposing a new development. Someone, of course, will then have to respond to that complaint. Will that really speed up our ability to build housing?
You might counter that A.I. will solve this problem by quickly summarizing complaints for overwhelmed policymakers, much as the increase in spam is (sometimes, somewhat) countered by more advanced spam filters. Jonathan Frankle, the chief scientist at MosaicML and a computer scientist at Harvard, described this to me as the “boring apocalypse” scenario for A.I., in which “we use ChatGPT to generate long emails and documents, and then the person who received it uses ChatGPT to summarize it back down to a few bullet points, and there is tons of information changing hands, but all of it is just fluff. We’re just inflating and compressing content generated by A.I.”
also interesting:
Our societywide obsession with speed and efficiency has given us a flawed model of human cognition that I’ve come to think of as the Matrix theory of knowledge. Many of us wish we could use the little jack from “The Matrix” to download the knowledge of a book (or, to use the movie’s example, a kung fu master) into our heads, and then we’d have it, instantly. But that misses much of what’s really happening when we spend nine hours reading a biography. It’s the time inside that book spent drawing connections to what we know and having thoughts we would not otherwise have had that matters.
“Nobody likes to write reports or do emails, but we want to stay in touch with information,” Mark said. “We learn when we deeply process information. If we’re removed from that and we’re delegating everything to GPT — having it summarize and write reports for us — we’re not connecting to that information.”
I don’t understand the logic of this text; who’s saying that A.I. is supposed speed up our ability to build housing? The NIMBYs are obviousely trying to use A.I. to slow it down. The legislators make it possible for A.I. generated complaints to slow it down and they or their puppetmasters may also have an interest in slowing it down.
You might counter that A.I. will solve this problem by…
What problem? Put in another (Matrix) way: “there is no spoon”.
Jonathan Frankle, the chief scientist at MosaicML and a computer scientist at Harvard, described this to me as the “boring apocalypse” scenario for A.I., in which “we use ChatGPT to generate long emails and documents, and then the person who received it uses ChatGPT to summarize it back down to a few bullet points**
It is completely unclear to me if this scenario is supposed to be about the legal battles about building houses, or legal battles in general, or about all kinds of e-mails and documents. It seems kind of obvious to me that the network of people that basicaslly rule the world (leaving open the possibility that the world may allready be run by A.I. or reptiles ) can control to what extent this scenario may be possible; in the case of legal system regarding building houses the legislators may dictate that the complaints may only be a certain length (not a 1.000 pages) and that alone will greatly diminish the A.I legal warfare scenario.
Our societywide obsession with speed and efficiency has given us a flawed model of human cognition that I’ve come to think of as the Matrix theory of knowledge.
What on earth is a “societywide obsession”? How many people in our society have this obsession? If you get paid per brick as a bricklayer this obsession may take a hold of you, but if you get paid by the hour perhaps not so much. The management of companies may generally focus on speed and efficiency in the realm of a production process, but even here we have exceptions.
the government/company pays (another) company for a project itself and additional for all the worked hours, giving it not much incentive for speed and efficiency.
civil servants are known for being lazy; if bureaycracy was speedy and efficient it would be called something else.
I also don’t understand what model of human cognition is being talked about and who is supposed to have this model and how is this model is supposed to be the result of the above mentioned obsession.
Many of us wish we could use the little jack from “The Matrix” to download the knowledge of a book (or, to use the movie’s example, a kung fu master) into our heads, and then we’d have it, instantly. But that misses much of what’s really happening when we spend nine hours reading a biography.
This is a classic example of a bait and switch:
The bait is the book that people want to download (into our heads) rather than read (because it is not a fun book to read).
And then later in the text that book is switched for a book that we do want to read, namely a biography.
Furthermore the reference to the Matrix movie in which Neo get’s a kung fu download is shaky in my opinion in that even in the movie the kung fu skill completion requires a fighting simulation with Morpheus. This also gives rise to the idea that you can download a book in such a way that the " drawing connections to what we know and having thoughts we would not otherwise have" may happen after the download.
Nobody likes to write reports or do emails, but we want to stay in touch with information,” Mark said.
Some people like to do e-mails and write reports and some don’t, just like some people like to drive all day on a forklift and some would rather bake bread or pizzas.
…but we want to stay in touch with information.
Who want to stay in touch with information and what information do they want to stay in touch with? For what it’s worth: I know many coworkers that love to do e-mails, so they (at least for some part of the day) don’t have to do the other parts of their job.
We learn when we deeply process information. If we’re removed from that and we’re delegating everything to GPT — having it summarize and write reports for us — we’re not connecting to that information. ”
This looks to me as a perfect example of a “deepity”, a word popularised by philosopher Daniel Dennet, meaning a statement that sounds apparently profound, but is actually rather meaningless and sometimes even simply untrue. I mean, if you teach a child to change a lightbulb, how deeply does this information have to be processed for it to to qualify as learning (and what does “deep” even mean?)? Having a summary of a large (like a 1.000 pages) report may be just the thing you need to connect to that information. Furthermore this statement literally contradicts the early statement in the article:
there is tons of information changing hands, but all of it is just fluff. We’re just inflating and compressing content generated by A.I. ”
So again, a summary (created by A.I.) is what we need to connect with this kind of information as opposed to something that prevents that connection.
Anthropomorphism much? When did AI ever make a promise of any kind?
Besides, the opposite of Artificial Intelligence would be Natural Intelligence; not Human Intelligence. Otherwise, we need to stop talking about how intelligent dolphins, gorillas, etc. are… because if we don’t, a gorilla communicating with a human using sign language would also be ‘artificial’ intelligence, seeing how a gorilla is not ‘human’.
And neither ChatGPT nor GPT-4 is synonymous with AI either. Most these articles get stuck in some philosophical debate or science fiction scenarios whenever they want to, how should I put it, “do something with the buzzword AI”.
Russel and Norvig group intelligent agents into 5 categories:
Simple reflex agents
Model-based reflex agents
Goal-based agents
Utility-based agents
Learning agents
The first two for sure, maybe even the first three, wouldn’t be considered AI by a layman these days. Utility is to be understood as in the economics term and acts like a rational agent by maximizing its utility function; or, if you have to anthropomorphize it… maximize the agents ‘happiness’. There’s no learning required for it to act as an AIA (autonomous artificial agent).
However, this part of AI… nobody talks about. Instead, we like us some Learning agents, deep learning, large language models, that can make us feel like we have human interaction, so that we can start giving it human attributes in our minds… let it become self aware.
Now, I wish I could close by saying… why don’t we feel threatened by “artificially” intelligent gorillas using sing language to communicate with us, taking over the world? But then, there is the original Planet of the Apes and its remakes in recent years, so I guess we are scared of them as well.
I once read a book by Adam Fawer called the “The Einstein code” that was not about Einstein or any kind of code. However both Einstein (Moonwalking with Einstein) and code (Source Code) are powerful buzzwords, so there you go.
Well, the original English title was Improbable, so you need to take that up with whoever thought that Onwaarschijnlijk would sell less on the Dutch market. My guess, De Da Vinci Code probably did good sales numbers… so guilty by association. Strangely, in German they decided to go with Null (eng., Zero) for the title of the book. But that’s nothing you can blame the author for.
LOL, no doubt… no, but seriously… the guy who wrote the article (whole different can of worms) also has a podcast and just went over the same subject there for an hour. I mean, if I had to sum it up in three word’s “snake oil salesman”.
) can control to what extent this scenario may be possible; in the case of legal system regarding building houses the legislators may dictate that the complaints may only be a certain length (not a 1.000 pages) and that alone will greatly diminish the A.I legal warfare scenario.