> Remember the first time an autocomplete suggestion nailed exactly what you meant to type?
I actually don't, because so far this only happened with trivial phrases or text I had already typed in the past. I do remember however dozens of times where autocorrect wrongly "corrected" the last word I typed, changing an easy to spot typo into a much more subtle semantic error.
Sometimes autocorrect will "correct" perfectly valid words if it deems the correction more appropriate. Ironically while I was typing this message, it changed the word "deems" to "seems" repeatedly. I'm not sure what's changed with their algorithm, but this appears to be far more heavy handed than it used to be.
SwiftKey has this really frustrating one where you'll remove the incorrect word and try again, and it reinserts the wrong word plus something additional.
Not from traditional auto complete, but I have some LLM 'auto complete'; because the LLM 'saw' so much code during training, there is that magic that you just have a blinking prompt and suddenly it comes up with exactly what you intended out of 'thin air'. Then again, I also very often have that it really comes up with stuff I will never want. But I remember mostly the former cases.
I see these sorts of statements from coders who, you know, aren't good programmers in the first place. Here's the secret that I that I think LLM's are uncovering: I think there's a lot of really shoddy coders out there; coders who could could/would never become good programmers and they are absolutely going to be replaced with LLMs.
I don't know how I feel about that. I suspect it's not going to be great for society. Replacing blue collar workers for robots hasn't been super duper great.
> Replacing blue collar workers for robots hasn't been super duper great.
That's just not true. Tractors, combine harvesters, dishwashers washing machines, excavators, we've repeatedly revolutionised blue-collar work, made it vastly, extraordinary more efficient.
I'd suspect that these equipments also made it more dangerous. They also made it more industrial in scale and capital costs, driving "homestead" and individual farmers out of the business, replaced by larger and more capitalized corporations.
We went from individual artisans crafting fabrics by hand, to the Industrial Revolution where children lost fingers tending to "extraordinary more efficient" machines that vastly out-produced artisans. This trend has only accelerated, where humans consume and throw out an order of magnitude more clothing than a generation ago.
You can see this trend play out across industrialized jobs - people are less satisfied, there is some social implications, and the entire nature of the job (and usually the human's independence) is changed.
The transitions through industrialization have had dramatic societal upheavals. Focusing on the "efficiency" of the changes, ironically, miss the human component of these transitions.
> Remember the first time an autocomplete suggestion nailed exactly what you meant to type?
I actually don't, because so far this only happened with trivial phrases or text I had already typed in the past. I do remember however dozens of times where autocorrect wrongly "corrected" the last word I typed, changing an easy to spot typo into a much more subtle semantic error.