It’s not lying or hallucinating. It’s describing exactly what it found in search results. There’s an web page with that title from that date. Now the problem is that the web page is pinterest and the title is the result of aggressive SEO. These types of SEO practices are what made Google largely useless for the past several years and an AI that is based on these useless results will be just as useless.
Yes, that’s what I said. My native language is a romance language too. And after speaking it her whole life, my wife has trouble getting the grasp of how in English swapping two words completely changes the meaning of what she’s saying (especially when it’s two nouns, like e.g. “parent council”)
The thing is that in French, Spanish, etc. it still makes sense if you put the adjective before the noun, even if it might sound weird in some cases. An adjective is an adjective and a noun is a noun.
But English is positional. Where you put a word gives it its function. So “red car” and “car red” mean different things.
open_dialog_file or dialog_open_file?
Tech Leads and Staff+ Engineers are still IC roles. If you’re not managing people, then you’re not in a manager role.
Yes, but 15 spelling variations doesn’t mean the name wad spelled 15 times in total (as the comment I replied to says), because each variation can be used multiple times.
Ok, I now see the discrepancy in the article (“misspelled 14 different ways” and “spelled 14 different ways”). I was confused by the commenter’s phrasing (“spelled 15 times in total”)
How are crashlytics and firebase analytics profiting off of users? I cannot imagine not including those in an app you’re actually hoping to improve.
No it doesn’t. He could have his name show up 100 times in the documents, 30 of which are misspelled. The 30 misspelled occurrences would show 14 unique spelling variations.
Edit: I see what you mean now. He spelled his name in 15 different ways if he misspelled it in 14 ways (unless he never spelled it right)
Even if you’re looking at a range, it still won’t tell you anything except that you found the guy or you didn’t find the guy. If you didn’t, what’s the next step? (In the find when the suspect passed in front of the camera scenario)
You seem to talk about different things when you say “visual clue”. Yes, there will be a small duration in the video where the event happens and maybe a short aftermath. That’s not a visual clue, that’s the thing you’re looking for. What all others mean by visual clue is a definite indicator that you can see when picking any random frame in the video that tells you if that frame is before or after the event. That allows you to exclude all other frames from your search, reducing your search range by half.
A stolen bike, a broken window, your examples that trash the place or end up with a crowd of people in the area, all leave such a visual clue. At any random frame you can check if the bike is there or not, the window is broken or not, etc.
But let’s say you have footage of the street facing CCTV and you need to find at what time the suspect left the scene (crime happened somewhere else). There’s nothing that tells you when looking at the halfway point if the suspect already passed or didn’t. You still have to look at both sides of that point in time.
The classic example for binary search is looking for a word in a dictionary. You open it halfway and see if the words there are before or after the one you’re looking for. Then you know which half of the dictionary you need to look in next. Then you use the same method for that half and so on.
But what if someone highlighted a word in the dictionary and you don’t know which word? Binary search is useless. You have to skim through the whole thing until you see it.
Maybe I have no understanding of what a binary search is.
No, you’re not the one who has no understanding of what binary search is.
It probably depends a lot on where you live. My wife’s bike got stolen and she was woken up by police coming to check on it (one of the maintenance guys at our apartment noticed a man at 7-Eleven riding it and recognized it; came back running to check if it’s indeed missing and called the police). We fully expected the police would do nothing about it (it was the cheapest Walmart bike), but an hour later they called that they found the bike and have the culprit in custody. It did help that the bike was a girly mint green with a wicker basket, so they instantly recognized it when they saw it.
Then again, in San Francisco, when my wife got her car window smashed and wallet stolen (she was late for class and dropped her wallet under the car seat, didn’t stop to take it; but it wasn’t the wallet that caught the thieves’ attention, it was the breast pump bag that looked like a laptop bag; they threw it on the floor when they saw what it was), we never heard anything back from the police.
As a vim user myself, I don’t understand why you need relative lines either. I can just as easily type :23 to go to line 23.
I use it both ways. As a software engineer I use it for various packages, which don’t even need a releases page. But also as an end-user of open source software, I use it to download pre-built binaries of said software. Idk if you know, but there’s a lot of open-source software out there. And github is the most popular platform for hosting it. And when I say software, I mean the kind where you don’t expect your users to know how to build it from code themselves.
Why would your company use that? Did they use github for public applications targeted to non-techincal users? Because that’s what that page is for and what a huge chunk of Github users do.
So when you just needed software to run on your machinr, you built it yourself. But first read every single line of code to ensure that it’s safe. Did I get that right?
Because if you don’t trust the developer to provide safe binaries then you wouldn’t trust the same developer to provide safe code either.
If you use it as a developer you don’t care about the releases page. You want to see the code and for latest version you just need the git tags. But I’ve also used it for stuff I just needed to run on my machine as an end-user. And for those you turn to the Releases page. That’s where pre-built binaries go.
But it also depends on the target audience. Some projects, even if meant more as software to run than code to import, still target mainly developers or tech users in general and will not have more than just instructions on how to build them. Others, say a Minecraft launcher, or some console emulator, will target a wider audience and provide a good Releases page with binaries for multiple platforms.
Deep learning did not shift any paradigm. It’s just more advanced programming. But gen AI is not intelligence. It’s just really well trained ML. ChatGPT can generate text that looks true and relevant. And that’s its goal. It doesn’t have to be true or relevant, it just has to look convincing. And it does. But there’s no form of intelligence at play there. It’s just advanced ML models taking an input and guessing the most likely output.
Here’s another interesting article about this debate: https://ourworldindata.org/ai-timelines
What we have today does not exhibit even the faintest signs of actual intelligence. Gen AI models don’t actually understand the output they are providing, that’s why they so often produce self-contradictory results. And the algorithms will continue to be fine-tuned to produce fewer such mistakes, but that won’t change the core of what gen AI really is. You can’t teach ChatGPT how to play chess or a new language or music. The same model can be trained to do one of those tasks instead of chatting, but that’s not how intelligence works.
Any type of content generated by AI should be reviewed and polished by a professional. If you’re putting raw AI output out there directly then you don’t care enough about the quality of your product.
For example, there are tons of nonsensical articles on the internet that were obviously generated by AI and their sole purpose is to crowd search results and generate traffic. The content writers those replaced were paid $1/article or less (I work in the freelancing business and I know these types of jobs). Not people with any actual training in content writing.
But besides the tons of prompt crafting and other similar AI support jobs now flooding the market, there’s also huge investment in hiring highly skilled engineers to launch various AI related product while the hype is high.
So overall a ton of badly paid jobs were lost and a lot of better paid jobs were created.
The worst part will be when the hype dies and the new trend comes along. Entire AI teams will be laid off to make room for others.
I was confused why a package manager would need to import posts from a social network.
Why name a new product the same as a very popular existing product?