Also the outtakes clips (under the credits) were hilarious. youtu.be/2h0n3_ZIqt4

After watching the seasons, I read book 1 and decided it was close enough to season 1 that I could jump to book 7. I'm sure books 2-6 have lots to offer over the series, but I didn't have the patience to go through them before getting to the new stuff (which I turned out to like a lot).

Tip: If you end up watching all six seasons and want more, you can read books 7-9 w/o reading books 1-6.

We lived on a small street that had a slight hill leading to a long straight street that had a big hill that ended in a heavily-traffic'd street. I imagined sliding down the long street and then getting T-boned. So I tended to take it at <5mph but even at that speed you could start slipping.

When I taught my kids to drive after a snowfall on a snow-covered quiet neighborhood street with no one around at a very slow speed I had them jam on the breaks to let them feel what it's like to be in a multi-ton object that they have no control of on the theory that the feeling breeds caution.

I don’t disbelieve you (i.e., that this is true for you) but I want to say that as someone who’s written computer software for 50 years, the boost I can get with AI-assisted software development tools is truly amazing. I think people in other fields can or will see big benefits too.

I don’t know…this “insurance” thing sounds a lot like communism! :-)

But the end result--the behavior and capabilities of a gen AI system vs. a search engine--are pretty striking.

Gen AI is "just" a refinement of what search engines have done for a long time--trawl over a huge body of data, store it in some sort of data structure suitable for querying, and build a mechanism that does those queries.

Finally, I'll pull out one of my favorite phrases: Sometimes a difference in degree becomes a difference in kind.

Presumably evolution has endowed the human brain with all sorts of capabilities beyond "prediction" and I imagine whatever the gen AI builders are doing when they talk about adding "reasoning" to their models is analogous. TBD how successful they'll be.

True but (this must be a cliche by now but I'll say it anyway) so, to a real and relevant extent, is the human brain. Maybe not the simple word-follow-word way, but still I think the "just a prediction machine" isn't the best way to call out the limitations of gen AI.

After 25 summers in RI I've finally internalized that there's a Monday in August that I can not go to the dump and I make sure I know which Monday that is!

And AI doomerism feels a lot like various "end of days" cults / thinking in the past.

On Thursday I spent 45 minutes on 138 on Jamestown and the Newport Bridge heading to an 11am concert at The Elms. Fortunately I'd happened to look at Google Maps well before I'd planned to leave and left early. Got there at 11:03 and the performance started one minute later. An RI thing, sort of :-)

I think in most printed material (magazines, books), at least for the past 50-100 years and at least in the US, the standard layout would have little or no whitespace around em dashes. But feel free to be a rebel! :-)

No spaces around those em dashes though—please! :-)

Similarly, this one from Gmail. To import a file that contains filter descriptions, you first have to click the button labelled "Choose file" (which brings up a file-open dialog) and then click the button labelled "Open file".