A.I. still can't do basic bitch things

Cool photos you have there Bing, holy shit you suck and you are the best image search there is. Google is retarded enough to confuse niggers with gorillas, Bing thinks detailed picture is photo, Yandex classifies niggers as white. Are the commonly available training data sets these corporations use deliberately sabotaged?

Attached: bing photos.png (623x637, 542.9K)

user artists and programmers will be replaced by AI any day now I swear

...

As far as I know they don't use "A.I" for search engines and simply use keywords generated in various ways (filenames, alt property, text featured on the page of the image, embedded tags inside the image etc).

No, they do actually use various image recognition & categorisation techniques. Not just tagging and whatnot. But "AI" is a meme - it's really all just a variety of statistical techniques and doesn't "reason" or anything like that the way humans do, it's just a massive set of probabilistic models basically. And it's massively impacted by the training data & how specific it's trained to be.

SJWs in Silicon Valley get butthurt that if you train a model to say "What does a typical criminal look like", and it spits out a Black guy - the thing is though, that's just what the raw data says. Now, there can be all kinds of reasons, including "muh evil White patriarchy Capitalist cis oppression", but it doesn't change the fact that's what the raw data being fed to the model says.

So now they're obsessed with "removing bias" ie, massaging the data to fit their own political model. So the results are actually getting worse over time, which is hilarious. But without doing that "removing bias" step the Tay effect inevitably happens - absent massaging the training data all "AIs" inevitably turn into the average Zig Forums or /b2/ shitposter, even without actually being trolled. The raw data of measurable, objective reality itself causes it. Even Watson (the IBM system) had the same issue, although it wasn't public knowledge (partly why it's been massively gimped as well, same as the other chatbot engine APIs).

In conclusion, the reason AI thinks nigs look like gorillas is because they often do look like gorillas, not because of "muh rayssisms". You could probably create a dedicated set of models specifically just to pick out nigs from gorillas, and I imagine that's the direction they'll need to take.

AI is almost certainly just a plausible deniability cover for public policy fraud.
There is no way SiliCON valley has any honest intentions at all.

What do you mean by "public policy fraud"? I work in the field - feed raw data to a model you train yourself and you get exactly the same results. The fraud occurs because you can fake training data - but that's all happening in the direction of "equality".

Niggers are categorically not gorillas. Anything to the contrary is insulting to gorillas.

You detect fake training data or a skewed model by running your own model and noting discrepancies. Now imagine you couldn't get real data because the public data was faked.
Any models you trained would be wrong and you'd have no idea. This would enable ridiculous levels of fraud and that's why tech faggots are so excited.

I often confuse them too, and I'm not retarded.

ai isn't dumb, it's being made dumber on purpose. try and search anything on google and see what you get, really. or better yet, something more measurable, just look at jewtube homepage suggestions lately. the site just shits itself and doesnt have anything to suggest that's relevant and degrades into normie bullshit. there's actually a bottom of the homepage, you can reach it (i would know, i hit it around 20 times a day trying to get a single video worth even clicking on for a second). and if you watch anything slightly off color, prepare to mysteriously have your ads stop displaying for the rest of eternity (this is on the tv app too so no adblocker, you'd think if you're watching something on tv the ads would be worth much more, too). apparently if you're not an idiot, the site doesnt even bother advertising to you, like the ai knows every ad is for something completely worthless across the WHOLE site. the dumbest gorilla nigger design in history.

>"oh i see you've blocked 50 times, are you suuuure you don't find funny? here, i'll suggest it again"

>"you sure watch a lot of (((negative))) videos, user. how about a category to revitalize your faith in humanity?"


if google let its fucking youtube search/homepage ai off the reigns for five seconds the entire population of earth would gain 20+IQ overnight. they could even manually block anything regarding the jews in the algorithm, it wouldn't matter if every nigger was shown videos about compiler design and physics instead of the latest "purple drank, dat lean yo.mp3" music video.

AI really relies on massive amounts of data and anytime that fails, it's useless or worse than useless. Google Translate, despite access to so much data, still sucks.

hehe
valence baby

AI would have to have sentience to replace those two fields.

Not in case of artists because they arguably have no sentience, they have been recycling the same tropes for thousands of years.

influenced, biased, tuned, predetermined manipulated, selected by the ministry of wrongthink, thinkpol, propaganda, totally normal and entirely neutral group of editors.

No they haven't unless you're talking about "modern art"

there is no ai. its all just preprogrammed things that the computers do.

But you are a brain, a neural network greater than anything Google has ever produced. Can you not tell the difference yourself?

...

It won't matter. If I start documenting or noting reality so you can get a better set of data I'll just be banned. This already happens and we haven't even begun the dark age of tech priests.

You are observing a mismatch between fidelity vs. A.I. construct. Almost none of these systems are A.G.I. or strictly NN. They are more like 3rd gen systems similar to Watson, Siri, or Alexa. If you would like a more true experience you should try out 6th generation A.G.I. from a research institution/college.

Nigger, when will you stop posting word salads pretending your some big shot in whatever the thread is about?
You aren't funny, you aren't "trolling" anyone, you're just sad.

Hmm sir, you may have a point there, but gradient descent on heuristic fluctuations and cNN radian Markov constructs prohibit exact Z valuing.

Attached: 87b829d3889a3a19901c0b29d54e3d663a64faba21bd5e3aeb7bcb34b3bacb85.jpg (400x519, 19.96K)

talk science to me :^)

Kek got em there user

Bing's photodna and search is so superior that it can spit out nudist lolis

Prove it.

When the day comes that AI can replace programmers, that means meatbags are obsolete and will be wiped out and replaced by superior synthetics.

When I search something specific on Google, all I get are fucking retarded news articles. I can't be the only one noticing this? It's not like that for Bing.

The core problem I see is that AIs don't learn how to rank relevancy/attention to certain patterns, when another group of patterns are already detected (or not). To put it another way, they lack the ability to learn how to filter and regulate perception to learn the true boundaries of any task. That's why they need so much data and are so brittle, because the input is processed as everything at once. The reality is that it needs a "wandering eye" method instead, breaking down perception into smaller incremental (but extremely fast) steps to build a hierarchy of learned behaviors. Something like that would require inhibitory feedback loops in neural networks to manage it.

This is important as managing coincidences being learned by perception context, would go a long way to improve AI. As a side effect it would outright solve their "bias" problem and then some, without needing ridiculous amounts of data.

Infinity divided into infinite seperate parts for a reason.

OP IS A LOSER
SKYNET IS COOL

Ai will be a guy in an office making the source of your own pissed off mentality. Best bet cop out and corpratize your ideologies. Only in self sufficiency can any population remain in the eyes of the controller. No different now then as we have been in the past.

That's why I'm so excited to be a pioneer of AI automated mind control against my free will. Every day is an exciting adventure of how it's going to fuck up my life.

Instead if just doing what I want, it has to make my ears ring for about 30 minutes before doing what I did before.

It's like downloading your next behavior as opposed to just doing it.

The best part is that the faster it downloads the more it damages your brain and pisses you off from irritation.

I like how people like you decided to make AI control people's behavior based of data collection on the internet when SOME OF US had a completely different persona irl

Well to be fair, there is hardly a difference.

I was actually the complete opposite person of my online persona. Now I have an AI bossing me around IRL to fit my online data.

It's the most inhumane and stupid thing I've ever seen

modern "art" is an attempt to break away from those tropes that failed miserably and turned into a money laundering scheme

Imbecile, look at the image.

Actually it's about restricting to specific input nodes (rest are inhibited initially) to learn partial states of input. Then after it has processed that, it randomly inhibits different input nodes to repeat the process (doesn't need to be only input nodes, any node on any layer could be targeted). The idea is that the inhibitor feedback part is trained to find correlations between sets of partial data. Then eventually it learns how to purposely select which input nodes to inhibit next, upon exposure to certain data prior. Much like how a person's eyes will rapidly shift focus towards specific points of an image, based on what they initially recognize first.

You can immediately see the advantage here if it's practical to implement in deep learning. Because instead of needing to focus on literally everything in an image (all the time), it'll eventually learn where it should look and by how much, based on an initial patterns/states it notices first. More importantly, it'll learn what to ignore as well.

Attached: 34c7c53a72e628202a9f64b696dc4567d3df9f5818132e5f327678d32ac81d4f.jpg (1861x1341 111.33 KB, 211.3K)

I'm looking. It has photographs selected as image search result type yet the dumb ass bing is displaying paintings.