I'm not sure whether it's just the o4-mini which is failing this task for me or what, but it did not perform well on the pictures I provided. I took a screenshot of the photo both the times to avoid any metadata input.
E.g. I first gave it a passage inside of Basel Main Train Station which included a text 'Sprüngli', a Swiss brand. The model got that part correct, but it suggested Zurich which wasn't the case.
The second picture was a lot tougher. It was an inner courtyard of a museum in Metz, and the model missed right from the start and after roaming around a bit (in terms of places), it just went back to its first guess which was a museum in Paris. It recognized that the photo was from some museum or a crypt, but even the city name of 'Metz' never occurred in its reasoning.
All in all, it's still pretty cool to see it reason and make sense out of the image, but for a bit lesser exposed places, it doesn't perform well.
> This was probably an okay idea terribly implemented. GenAI creators on social media kind of sense.
It boggles my mind that there are people who think this is a good/ok idea. From a human perspective, all it does is pulls the mind ever closer to fictional imaginative world rather than encouraging real life interactions which I believe is inherently wrong no matter what business strategy is wrapped around it.
I recently arrived in Germany and movie/cinema aggregation is a huge issue and a hassle tbh, so I'll be using this frequently.
A suggestion, if I may, is to add language of the movie if possbile? It'd be great if there's any way to fetch that info and display it directly on your website instead of visiting every multiplex website to check it.
I don't think Bard can even perform at GPT 3.5 level, at least, not in my experience. Bard hallucinates pretty bad and you can't really be sure whether it's correct or not. Now, that's a given with any LLM, but in my experience, Bard usually gives you at least 1 wrong sentence in any of the results it produces whereas GPT 3.5 is mostly correct at surface level questions, but tends to hallucinate if you try to dig in deeper.
But, Bard can give you images as an output as well with the links to those images which may or may not be the correct links and GPT 3.5 can't do that. Still, in any general case, I'd say GPT 3.5 is way more reliable than Bard.
Yes exactly. It won't take long till someone exploits it which would be more robust than the prediction models we have right now and that's what I don't want a single dime of.
I know we're not there yet, and perhaps we won't ever be, or we'll be there in 10 years, but we don't have to map every single thing in the brain in order to exploit the foundational behavior.
> I dunno, I think I quite liked the world how it was before that!
I agree. I don't know why but when I read this article and the responses here stating how things can be related and can be measured, I suddenly felt this sigh? disappointment? for 'demystifying the mind'.
> Modern graphics looks better when you put them side by side, but the graphics 10 years ago was good enough that you easily get used to it.
Can you get used to it? Yes. Eg. Skyrim is still atmospheric, but graphically it is nowhere near what today's games offer. I'd say that one don't even have to look at games that are 10 years old and modern games side by side to see the difference. Someone getting used to older graphics is that person adjusting themselves to what they're presented, and not an objective observation and games today objectively look better even when they aren't compared side by side in my opinion.
> They feel pretty much interchangeable to me when it comes to graphics.
> The difference is nights and day...and I don't care one bit.
You used these 2 sentences back to back.
Tbh it's just you not caring about graphics whole lot, but can see the difference. To say that they're interchangeable is kinda hilarious because they're objectively better on PC, consoles than on Switch. You, yourself said so and the whole point in this thread was that of "how graphics definitely improved over the last 10 years" and not a subjective opinion of you not caring (which is fine but invalid in this conversation). Also, one can easily notice graphical differences between games that are released now and 10 years ago.
While I agree that there is somewhat over-reliance on CGI in movies nowadays, this post is like cherry picking the examples. Comparing bad CGI shots to few of the greatest films ever made (and with high budget for the time as well) just seems odd. I wouldn't say CGI itself ruined the movies, but how it is now implemented. There are films like Gravity which in my opinion, blended the CGI perfectly with what was happening in the scene and that's where CGI enhances the experience rather than otherwise.
Rare did go down after acquisition, but what others are you talking about? Their biggest (Zenimax) one will start releasing new Xbox and PC exclusive games from this year onwards, so it's still early to call that one a failure.
Mojang, is doing Mojang thing i.e. Minecraft and while they haven't had a great success after that, they're still doing fine.
Playground games is doing Forza which is actually pretty popular among racing players (perhaps the most popular).
Obsidian haven't started to make Xbox and PC exclusives yet, but they'll also start with that presumably next year.
I don't know where you're getting "Most of Microsoft's acquisitions in recent years have been failures." this.
On the other hand, Sony's way of buying studios is way different. Most of the times, they have already worked with those studios in past to make a PS exclusive and then they buy them. I agree that Insomniac is doing great (better than any of Microsoft's purchases), but other than that, there isn't much just like MS.
I think we will get a better picture of whose acquisitions work best after 5-6 more years.
You're comparing 2 games which were a generation of console apart and a lot of things happen between those 7-8 years.
> People will say, ah yes but the story, the lore! But if I wanted that I'd read a book.
That story and lore in the video game are exactly the reason I play video games i.e. video game can easily transport you into a world which you can see/hear/play-in instead of only imagining it in your mind. Books can never come anywhere close to that because of obvious restrictions of the medium and that was the reason, many (including me) were instantly captivated by the world of Skyrim. They successfully created a world full of atmosphere and lore along with the freedom of exploration and managed to do it excellently even compared to most of the games released now (and the thing is I also played Skyrim late in 2019, yet it is probably my 2nd or 3rd favorite game of all time).
It's unfortunate that you didn't like it, so have fun with RDR2 (which is an excellent game with arguably the best storyline and characters in any video game till date).
Not the OP, but in India, the lowest priced one I could find was from ASUS, which costs around $2600. I don't think the vanilla RTX 3080Ti is available here yet. The most expensive one was from MSI which costs around $3200. Whereas PS5 (official price i.e. without considering the shortages) costs $675.
E.g. I first gave it a passage inside of Basel Main Train Station which included a text 'Sprüngli', a Swiss brand. The model got that part correct, but it suggested Zurich which wasn't the case.
The second picture was a lot tougher. It was an inner courtyard of a museum in Metz, and the model missed right from the start and after roaming around a bit (in terms of places), it just went back to its first guess which was a museum in Paris. It recognized that the photo was from some museum or a crypt, but even the city name of 'Metz' never occurred in its reasoning.
All in all, it's still pretty cool to see it reason and make sense out of the image, but for a bit lesser exposed places, it doesn't perform well.