Spring is sprung
In a friend’s garden yesterday.
Quote of the Day
”At 50 everyone has the face he deserves”.
- George Orwell, in his notebook, 17 April, 1949
Musical alternative to the morning’s radio news
Carole King | Chains
Thanks to Doc Searls (Whom God Preserve) for the reminder.
Long Read of the Day
Artificial Intelligence and The Best Game in Town: Or How Some Philosophers, and the Behavioral and Brain Sciences journal, Missed a Step
Intriguing long read by David Lobina on the implications of the fact that LLMs (large language models) are essentially statistical engines. Includes his (rejected) response to a challenge issued by BBS.
Modern AI [i.e. machine-learning] is, in a way, glorified statistics, in the same way that a flat white is a glorified caffellatte – actually, a flat white is a badly made cappuccino, but the analogy still holds up: a former friend once joked that Mitt Romney seemed like a botched connectionist attempt at building a president. And as with Romney, probably, there is really nothing there; a chatbot such as ChatGPT doesn’t “know” any language, or any aspect of language, and it doesn’t “know” how to reason, either, certainly no more than a calculator “knows” how to, er, well, put two and two together (maybe the calculator understands the metaphor, though).
Indeed, and as stressed last month, deep neural networks connect an input with an output on the basis of the gigantic amounts of data they are fed during so-called “training”, when the relevant correlations are calculated. In the case of LLMs such as ChatGPT (note that I’m conflating an LLM with the “dialogue management system” that queries an LLM; will come back to this), the model predicts one word at a time, and only one word at a time every single time, given a specific sequence of words (that is, a string of words), and without making any use of the syntactic or semantic representation of the sentences it is inputted.
That is, LLMs calculate the probability of the next word given a context…
Interesting if you’re fascinated by the fuss about LLMs. But it may also be an acquired taste.
One of the things that’s very interesting about ChatGPT is the way programmers are using it. I’d been wondering about this for a while (and indeed in the first piece I wrote about it had likened its arrival to the arrival of VisiCalc, the first spreadsheet in the late 1970s) but this post by Dave Winer (Whom God Preserve), arguing that “ChatGPT clearly has a place” elegantly crystallised that view.
This thread is worth money. I’ve given ChatGPT programming jobs like the ones the author describes, and it’s saved me huge amounts of time. Last one was asking how to do something with the Twitter API. I could have spent fifteen minutes trying to find it in the docs, or on Stack Exchange, but I got the answer instead in a few seconds, and there was no bullshit, no preambles, just the answer to the question I asked.
Journalists, who do most of the writing about news, immediately focus on how it might affect their careers, and imho educators zoom past the purpose of education, to create more better-educated people. As a kid, I had a party the day my parents bought us an encyclopedia. That meant we could settle arguments by getting facts. We could’ve gotten them before but that would’ve meant a trip to the library. Better tools make for better information. ChatGPT is a revolutionary tool. Kind of like Alta Vista was when the web first came out. I’m sure people screamed that it would screw up something. People always say that about change, esp people who are invested in the way things are.
Maybe there will be negative consequences of ChatGPT, but I’m sure we’re not in a position to see what they are now, based on experience with similar changes. And maybe we’ll look back on this moment twenty years from now, and not be able to imagine what life was like before we had this fantastic tool.
My commonplace booklet
Her car died, so she walked to work. One day on the walk, she found $15,000.
If you’re overwhelmed by the awfulness of the world, here is a heartwarming story from the Seattle Times.
This Blog is also available as a daily email. If you think that might suit you better, why not subscribe? One email a day, Monday through Friday, delivered to your inbox. It’s free, and you can always unsubscribe if you conclude your inbox is full enough already!