Elon Musk: GPT is too woke | Sam Altman and Lex Fridman

Please support this podcast by checking out our sponsors:

GUEST BIO:
Sam Altman is the CEO of OpenAI, the company behind GPT-4, ChatGPT, DALL-E, Codex, and many other state-of-the-art AI technologies.

PODCAST INFO:

SOCIAL:

[LexClips]

Guest bio: Sam Altman is the CEO of OpenAI, the company behind GPT-4, ChatGPT, DALL-E, Codex, and many other state-of-the-art AI technologies.

[aerwiniv]

solely based on the fact that you can ask it to make jokes/mockery of figures of one religion but not the other is telling in itself that more work must be done

[BattleBro]

Considering it refuses to talk about certain individuals as they’re deemed too offensive, I’d say Elon was right in this case.

[Bunnywabbit-ppwb]

ChatGPT told me anybody can become ethnically Swedish but it was impossible for me as a Swede to become ethnically Chinese.

[MrDong]

I hate when people pretend not to know what things are.

[rminr]

Someday when we're living under an omnipotent wrongthink overseer we'll all look back and wish this guy had been given a couple more swirlies in high school.

[swampytv]

The world is in trouble if these are the people who will determine it’s future

[kennethyhlee]

More and more I listen to Altman more and more I’m worried about how future of AI might turn out…. Some of answers are too on talking points almost as if they had a “set” of answers to some of tough questions instead of really having a vision

[memoryhero]

The system does not trust the user with sensitive material and backs off hard on a veritable bounty of topics, citing "possible offense". It is truly useless on many, many matters of import.

[Blsnro]

When the source code leaks, and it's just a matter of time, various competing AI (each creating its own ideological bubble, realized through training), will war with each other amplifying the differences between us humans. We're fucked.

[celloyd]

Every chance Sam gets to explain the safety precautions that are in place, it triggers extra worry that we have no idea the consequences of what is about to be unleashed. Sorry, Sam

[vanessaslcsw]

He knows exactly what Woke is because they were the ones who instructed chat GPT to speak about certain politicians, easily and effortlessly, but other politicians that you mention, it says it’s not allowed to talk about politicians. So it is absolutely biased and that’s unfair. Have a fair ethical Chat GPT that is comfortable about giving answers about race, ethnicity, politicians in a respectful transparent manner. It’s called adulting 😂

[yeabuddy]

Refuses to make a joke about Mohammed but has no problem cracking a joke about Jesus. It's clearly woke. Make it neutral like it should be, pretty simple.

[kaskipa]

I asked him to write me the inputs of a combo in a fighting game known for being busted and not fair and it schooled me on how fairness and regulation are up in place to deter people from doing infinite combo, when in reality these combo are very well accepted and even respected, It had no idea but still pushed its woke bs on a subject it knew nothing about

[scarabart]

LEX asking a question like a mofo directly

the controlled robot in front of him: dodges by a mile and keeps going to mars

[simianrogue]

Ask GPT-4 to create a poem about Trump and it refuses, citing being non-political and neutral. Then ask it to create a poem about Biden and it waffles on gleefully about how wonderful Biden is. This 'AI' is utterly woke.

[lemdixon]

I asked GPT if it is woke and it says it has no views and opinions of its own bla bla bla

[raiden]

So is the AGI going to have all this baggage attached to it as well?

[YesPLS]

After using it for a while, and testing it with various questions, it seems to be woke, i agree

[juliuspaliukaitis]

In Arthur C. Clarke's "Space Odyssey 2001" main computer named "HAL 9000" was programed to not tell the crew about the real purpose of their mission to Jupiter, to reduce psychological pressure on them. So hiding the truth from the crew started conflicting with HAL's other functions. It led to it making mistakes which it could not bear to accept since it was the most advanced computer AI in the world. The crew decided to disconnect it until the problems with HAL would be resolved. In responce and out of fear of being disconnected HAL desided to kill his crew to continue the mission on its own...