@Scobleizer @LEGO_Group As a LEGO lover and collector, the ultimate dream is that a GPT model (maybe not GPT-4, but a future model) will be able to “look” at the PDF of a set (which @LEGO_Group provides online) and generate an alternative set of instructions to build anything you have asked.

AI-driven misinformation will just not target today. It will also target yesterday. Especially yesterday. https://t.co/dBu4lF7kdN

Told you. https://t.co/uiENyWcEgw

Wouldn’t it be amazing if @vivaldibrowser would team up with @StabilityAI to offer OpenGPT integrated into the browser in the same way PsychoBing is integrated into SpammyEdge?

[whispers]

What? OpenGPT is not out yet? Oh.

@GaryMarcus @WIRED Super fun. Well done both of you. Here, get a sticker.

Operation lipstick-on-a-planetary-pig has started.
REPEAT
Operation lipstick-on-a-planetary-pig has started.

(from https://t.co/gopXHoQhAb) https://t.co/xnfithGUkW

And where the difference between truth and lie will be able to destroy people’s life while the audience cares practically nothing about it.

We could even see a new form of blackmail coming out of this. Pay or we’ll inundate the internet with pictures of you doing X.

When we talk about AI-driven misinformation & manipulation, people think about situations involving super-high-profile individuals like this gentleman.

But that’s not where the damage will be. It will be done to very minor targets who will not be able to defend themselves. https://t.co/ifeOat57Wl

@AiMachina @BKanizay @nicholas_dimes @spinnybuns @WizardConNFT @mars_eve @AiJoe_eth @HPLoveBux Interesting. Are we certain that this is product placement rather than the result of an inconsiderate training set?

Can you share more about the methodology you used to uncover it?

I’m just trying to not jump to conclusions.

Yep. https://t.co/doUOR76Jnc