It seems to me there is a word or two missing between “rich” and “slowly”. If I read the whole thing aloud I cannot parse it into a sentence. Or the word “rich” could be removed. That would be clunky but at least grammatically sensible.
“Make data get smoothed out” is a very strange way of saying “smooths out data”
> The weird, rare, surprising patterns [that make data rich] slowly get smoothed out when an AI model trains on outputs from a previous model.
i.e., the patterns are responsible for making data rich, and they are slowly lost as each new generation model trains on the prior generation's output.
Or, if you'd prefer an analogy, we're using a copy machine to output new documents by taking the last copy spit out by the machine, adding some marks to it, and running it through the copier again. Over time, details present in much older copies blur and fade away in Nth generation copies.
It might be weird if you haven't read a lot of English. It's actually quite normal to say that process X is a way to make effect Y happen. "Makes your mout water" is more effective than "waters your mouth". "Makes your breath fresh and tolerable" is better than "freshens and tolerablerizes your breath". Etc.
Actually, what you are describing is what happens when LLM-generated prose cycles and then trains humans to use equally dull thinking.
And then there's trucks flashing an indicator to say it's safe to overtake if you're behind them. In the UK it's the nearside indicator, which makes sense: it's a bit like the truck is pulling over to let you pass. In Aotearo, it's often the off-side indicator, so you think the truck is going to pull out in front of you. I've never understood what the Aotearoa drivers are thinking there
All. The. Time. And I hate it. Imagine giving a customer a rebate based on buggy code. You fix a bug, the customer comes back and wants to check that the rebate was correct that last time. Now you have to somehow hard-code the rebate they did get so that your (slightly less buggy) code gives the same result. But hard-coding has the risk of introducing other errors on its own. Oh yes, and you've never enough time to do things properly because Customers (or maybe Management). A tangled mess of soul destroying lifeblood-sucking code and pressures ensues.
back working on my lighting desk, after a couple of years of hating it because the communications bus between the many different modules was flakey and so the whole thing wasn't fun to use. I bit the bullet last year and re-implemented everything with CAN-bus communications and it's actually fun to use now.
Current work has been improving boot time. Was nearly two minutes because of one board, and that's a long time for the lights to be out if you have to reboot during a show. I'd wanted to use buildroot to get a custom kernel that should boot much more quickly, but the buildroot learning curve was steep for me, particularly as I've no expectation of ever needing the knowledge again.
Independently but concurrently I decided I really ought to understand what all this AI stuff was about, for fear of getting left behind. That coincided with the release of opus 4.5, and holy heck has it made a difference! With a little guidance from me Claude got the buildroot environment working and the boot time down to less than 10 seconds. I've been _really_ impressed. I've had Claude write a few boring utilities that I could easily have done but Claude managed much faster and with less boredom on my part. Fortunately for my AI revolution I think I'm a better Business Analyst/writer than I am a coder, so it fits with my temperament.
This is one of the things I actually remember my mother saying. Festina lente [0]: Make haste slowly. I've always tried to stick to it because when I have I've found more to appreciate in whatever I'm doing (as TFA says)
Sadly, for some reason I now can't read slowly, which pisses me off. I and my partner read aloud together alternating chapters of a chosen book, and I love how get _much_ more out of the book than I would reading alone in a tenth of the time.
I've also found that some books seem written to be read aloud: the sentence structure and punctuation lends itself to easy reading aloud, whereas some books have really convoluted sentences with multiple parenthetical sub-clauses that are a real challenge to read aloud in an a way that's easy to follow. I've ended up so that normally try to write in a way that's easy to read aloud. I think if something's easy to read aloud it's going to be easy to comprehend when read normally. And Yes, I know that the sentence at the beginning of the paragraph probably doesn't match that.
reply