Kbin_space_program

Yeah and it unfortunately succeeded, so I’m surprised it is still around.

What the hell kind of crap is that writing?

Also, I’m surprised this is even this a thing, I thought he’d abandoned the boring company years ago.

It’s a core problem with image generator LLMs. For some fucking reason they seem to have fed them the content from sites that had a lot of porn. Guessing Imgur and Deviantart.

Literally the first time I tried to use MS’s image generator, was out with some friends trying a new fried chicken place and we were discussing fake tinder profiles.

So I thought to try it and make a fake image of “woman senuously eating fried chicken”.
Content warning, blah blah blah.

Try “Man sensuously eating fried chicken”. Works fine.

We were all mystified by that. I went back a few days later to play around. Tried seeing what it didn’t like. Tried generating “woman relaxing at park”.
Again, content warning. Switch to a man, no problem. Eventually got it to generate with “woman enjoying sunset in a park.” Got a very dark image, because it generated a completely nude woman T-posing in the dark.

So, with that in hand I went back and started specifying “fully clothed” for a prompt involving the word “woman”. All of a sudden all of the prompts worked. They fed the bot so much porn that it defaulted women to being nude.

A camp of concentration, if you will.

It was a work issue about a decade ago. Client wanted certain emails from automation to be masked as coming from him.

Most email boxes, including Gmail, didn’t have an issue. Outlook(the one that shipped with Office) laughed at it and displayed the original sender in giant bold letters.

Adapt to it as best we can. Minimize your use of fossil fuels, particularly Natural Gas(Methane). Get some books on farming to understand a worst case need to live off grid.

We’re past the point that we can go back. The glaciers on Antarctica and Greenland are in a self-sustaning melt cycle at this point.

Yay, does this mean that Google is going to stop saying the masked email address is the sender and hide the true email address?

You know, like MS has done for over 15 years now?

Please don’t create straw man arguments.

I never said cold turkey or abstinence. I said “no drug use in the free apartment.”

No. The rehab is free.
Province is even putting people up in apartments.

Only rule is: no drugs in the apartment and you enroll in a free rehab program. 80% are turning it down willingly.

Also, “usually trouble”.

Have rarely seen anyone have trouble crossing there. Unless it’s Tesla, BMW or Audi SUV, then just let them go pass, because they can’t drive anyways.

Sir Terry Pratchett.

A phenomenal author whose ability to weave a story is fantastic, but was also adept at writing in jokes and references that make re-reading the novels a delight.

For Canadians, Environment Canada has a fantastic app called WeatherCan that includes all of the usuals, has doppler radar for all of North America with the option for a one or three hour loop.

It also includes alerts from Environment Canada for any location you’ve added and links to the marine weather forecast site.

1TB can be Recommended Chrome Ram?

Fun fact time: in the 18th and 19th century Britain, postal service was fast enough to have Charles’s reply back the same day, with enough time for Edward to make up some fun British insult and write it down and send it off.

At what point is a sauce not a curry, soup or drink?

Bug your indian friends by calling Ketchup a tomato curry.

You’re building beautiful straw men. They’re lies, but great job.

I said originally that we need to improve the interpretation of the model by AI, not just have even bigger models that will invariably have the same flaw as they do now.

Deterministic reliability is the end goal of that.

It’s an exaggeration, but its not far off given that Google literally has all of the web parsed at least once a day.

Reddit just sold off AI harvesting rights on all of its content to Google.

The problem is no longer model size. The problem is interpretation.

You can ask almost everyone on earth a simple deterministic math problem and you’ll get the right answer almost all of the time because they understand the principles behind it.

Until you can show deterministic understanding in AI, you have a glorified chat bot.

I used to not do it, but now I do because it’s ridiculously nice to just to have it all ready to go.

Third option: get blankets one bed size up.

King size comforter on a Queen size bed solved the “who stole the comforter last night issue” for me.

Will increasing the model actually help? Right now we’re dealing with LLMs that literally have the entire internet as a model. It is difficult to increase that.

Making a better way to process said model would be a much more substantive achievement. So that when particular details are needed it’s not just random chance that it gets it right.