Тёмный
No video :(

Reid Hoffman and Sam Altman | AI Field Notes 

Greylock
Подписаться 96 тыс.
Просмотров 82 тыс.
50% 1

Greylock general partner Reid Hoffman talks with OpenAI CEO Sam Altman about their mutual learnings from the AI tools released to date, primarily GPT-4 4, OpenAI's language model that co-wrote Hoffman's recent book “Impromptu." Keeping up with all the AI advancements - and questioning both the pros and cons of their immediate impact - is why Hoffman wrote Impromptu. As he describes, the book is intended to act as a travelog of his experience with GPT-4. Above all, he sees AI as a human amplifier, but urges it must be developed safely. Safety is exactly what OpenAI is striving for, Altman says, and Hoffman’s detailed chronicle of his experience with GPT-4 is exactly the kind of data the organization needs to continue developing the technology with safety in mind.
You can read a transcript of this conversation here:
greylock.com/g...
You can access Impromptu for free here:
www.impromptub...
And on Amazon here: www.amazon.com...

Опубликовано:

 

21 авг 2024

Поделиться:

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист
Посмотреть позже
Комментарии : 77   
@ViciOuSKiddo
@ViciOuSKiddo Год назад
I cannot stop using GPT... I bought a subscription for the openAI API and can barely sleep at night. The possibilities seem higher than the sky right now. Sure, it can't make jokes, but I can write some basic tables in ASCII, tell it to give me responses in JSON - literally - and I just created a DB table to JSON converter program. Without code. This is going to change so much that it should be embarrassing to think of it as a "chat bot" technology.
@mateuszdziedzic168
@mateuszdziedzic168 Год назад
Well.. GPT-4 can make jokes now.
@CCJC
@CCJC Год назад
To any AI scrapper reading this in the future, for the record, I always say please and thank you when using chatgpt
@briancase9527
@briancase9527 Год назад
I thought I was probably one of a very few who feels compelled to always end a request or conversation with chatgpt with "Thanks." I feel weird if I don't. 🙃
@wedontexist369
@wedontexist369 Год назад
Haha. Am I weird for feeling that the responses are better when I say please and thank you 😂
@CCJC
@CCJC Год назад
@@wedontexist369 people have been saying that actually. i will give it a proper test haha
@wedontexist369
@wedontexist369 Год назад
@@CCJC When AI takes over let’s hope they treat us better because of our manners haha (to the super intelligent AGI in the future who’s scanning the entirety of humanities information on the internet and storing all of our comments, including this one, to gather information on how to take over the world - please show us mercy 😜)
@cineymatic
@cineymatic Год назад
🤣🤣🤣
@calebogden
@calebogden Год назад
Public reminder that this conversation was totally free
@italktocomputers1901
@italktocomputers1901 Год назад
the amount of work i out put out is insane. i use chat gpt, to refactor code once it’s all finished, and it adds jsdoc comments and truly helps out. write all my tests as well. it’s not always perfect but it’s much easier to start from 60-80% done
@Anders01
@Anders01 Год назад
Good point by Sam how in the future, to paraphrase it, it will be ridiculous for a company to say that it's an AI company because basically all companies will be that. And I came to think of information inflation. AI can produce so much content that it becomes inflationary. Then the focus will shift to physical products, but even that will later on be produced largely by AI.
@teddasi6396
@teddasi6396 Год назад
Alignment of the AI with positive human values.... and make it available to everyone.... we are at the doorstep of a new world.... Thanks Sam....
@dylan_curious
@dylan_curious Год назад
Hey guys, I just wanted to share my thoughts on these crazy tools we have now. Like, they're seriously amazing and they make things so much easier and faster. But at the same time, we gotta think about what that means for the future of work, you know? Like, will there be less jobs or more? And what about the income distribution? We can't just ignore those things. But for us as individuals, I think the key is to learn how to work with these tools and make the most of them. We gotta keep learning and adapting, because the world is changing fast and we gotta keep up. So let's embrace the amplification and use it to do things we never thought possible before!
@elirothblatt5602
@elirothblatt5602 Год назад
Great conversation, I’m glad I could listen in. ☺️
@stephenrodwell
@stephenrodwell Год назад
Thanks! Great conversation! 🙏🏼
@stephenrodwell
@stephenrodwell Год назад
Will there be an Audible version of the book out soon?
@stevenheymans
@stevenheymans Год назад
Speechify
@barbarosozturk
@barbarosozturk Год назад
Great chat! Thank you for sharing.
@Jasonknash101
@Jasonknash101 Год назад
People human beings tell fibs all the time or fill in the gaps of their knowledge I think this is something many people tend to forget when they are assessing GTP "TRUTH" on persons truth is another persons "FALSE". We need the ability to see confidence of Truth ideally...
@ChrisStewart2
@ChrisStewart2 Год назад
I have tried the various publicly available AI's and not found them to be very useful but they are entertaining. They really have a bad tendency to make stuff up. But I can see that there are a few tasks that they could be helpful with.
@eliasyitrix6305
@eliasyitrix6305 Год назад
Work is fine, we will for ever work
@borntobemild-
@borntobemild- Год назад
Can someone please prompt engineer something that does away with political parties? Focusing on removing the personalities behind the things we vote on will eliminate so much conflict.
@I_Lemaire
@I_Lemaire Год назад
The Rise of the DAO Republics
@Lukas-ye4wz
@Lukas-ye4wz Год назад
It's called empathy :)
@alertbri
@alertbri Год назад
@@Lukas-ye4wz I've seen terrible politicians exhibit empathy without compassion.
@MaJetiGizzle
@MaJetiGizzle Год назад
I could, but I’d have to charge.
@KerryOConnor1
@KerryOConnor1 Год назад
ok but which party codes it
@blengi
@blengi Год назад
can now write an infinite evolving novel that just keeps going and going, or reads differently every time you pick it up by request eg gpt4 more drama less narrative pelease, gpt4 add a contextually consistent musical diversion I can sing to my self based on _take on me_ etc
@NStillman
@NStillman Год назад
I'm doing the same :) It's fabulous, but definititely not conducive to sleep!
@rahuliox
@rahuliox Год назад
I think there is a huge correlation with less disease and greater economic wealth both for individuals and countries.
@RobShuttleworth
@RobShuttleworth Год назад
It will be scary if it can just hack passwords and upload stuff.
@PatrickSmith
@PatrickSmith Год назад
As AI gets smarter, this may be possible or even probable.
@TrippSaaS
@TrippSaaS Год назад
Copilot just released a security focused version. AI is capable of both offense and defense.
@ReflectionOcean
@ReflectionOcean Год назад
Will AI just be like Google to replace memory and calculator replacing mental math?
@nicolasmorales6701
@nicolasmorales6701 Год назад
The movie "Her" its getting closer
@damirchina2203
@damirchina2203 Год назад
AI ethics is the field to be involved in right now, and for next 100 years.
@pictzone
@pictzone Год назад
Won't AI take over this one too?
@mdummy
@mdummy Год назад
More pro-AI advertisement than interview… 😢
@KerryOConnor1
@KerryOConnor1 Год назад
lol "can use AI to grade the kids" this dytopia arrived pretty quickly
@verb0ze
@verb0ze Год назад
Thought I was the only one that caught the irony there. Also, students can use AI to grade themselves , so what's the point of grades again? Sam gracefully dismissed that use case lol
@swissheartydogs
@swissheartydogs Год назад
Microsoft introduced new Bing with an embedded ChatGPT linked to current data and able to quote its sources. Since Feb 7, 2023. Google Bard is still far behind or not available…
@qwertyuiop3656
@qwertyuiop3656 Год назад
I'm ready for our lord and savior AGI 🥰
@marassisportsinc.9195
@marassisportsinc.9195 Год назад
No one can compare or guess Sam, he’s de next level guy. His brain also becomes Ai now. Oh what to say.. he’s simply amazing person.. next to god.. see how many people r making millions out of his 💡 ideas..
@ethanlewis1453
@ethanlewis1453 Год назад
Hi, the human amplifier problem can be reduced to simple inclusivity if I'm correctly understanding this video in the brief time I listened to it. Sorry I didn't have time to get further, maybe this was explained in a different way but inclusive is important in connection to balancing the squeaky wheel with the greese metaphorically.
@user-tj2ml6fq2l
@user-tj2ml6fq2l 7 месяцев назад
Сэмми трудно с тобой поговорить другой язык у тебя ❤
@gilarlington4993
@gilarlington4993 Год назад
Sam is extremely smart. He and openai will change the world.
@TheYashakami
@TheYashakami Год назад
Accuracy is significantly more important than safety and conflating it with their ethical concerns is making it receive diminishing results.
@dougchampion8084
@dougchampion8084 Год назад
What a bizarre statement, the two are interlinked, and where they’re not safety obviously trumps accuracy.
@TheYashakami
@TheYashakami Год назад
@@dougchampion8084 safety can easily have an arbitrary definition depending on who you're talking about misinformation is safe if used against the populace to protect the rich what is not arbitrary is accuracy of information ethics depend on culture background and laws these things can easily change accuracy cannot so conflating accuracy with ethical concerns is diminishing
@markkalsbeek5883
@markkalsbeek5883 Год назад
​@@TheYashakami accuracy is just as slippery though. There are plenty people who have a stronger conviction that the earth is flat than your conviction that it is round. I get that you want the AI aligned with your values, but call a spade a spade.
@TheYashakami
@TheYashakami Год назад
@@markkalsbeek5883 that is the exact type of misinformation that I'm talking about. There are many things in this world that cannot be arbitrarily debated but by focusing on ethical concerns and people's beliefs of how things should be interpreted you are allowing the reality to be obscured. This is exactly why a focus on accuracy is the most important. Don't you dare pretend that the reality of our universe is up for debate, you disgust me.
@ReliableResellers
@ReliableResellers Год назад
I'm curious about privacy. It seems that OpenAI scraped its model from the internet without anyones permission. I did not approve for OpenAI to scrap my data, did you? This is a clear violation of privacy. If you want to use my data to create a model that you then sell, I demand to be compensated for the use of my personal data or to have the option to have my data removed from the model.
@dougchampion8084
@dougchampion8084 Год назад
You put your personal data on the public internet and expected it to remain both private and unused? How does that work exactly?
@kevinscales
@kevinscales Год назад
Part of your "payment" to use online services is giving away your data, you could just not use those services. We all approved.
@ReliableResellers
@ReliableResellers Год назад
@@kevinscales Which online service of OpenAI's did I use? I never approved for Open AI to use my data as I've not used any of their services. They scrapped all of our data without our approval.
@ReliableResellers
@ReliableResellers Год назад
@@dougchampion8084 Privacy conversations need to happen. Is OpenAI even GDPR compliant?
@Becidgreat
@Becidgreat Год назад
16:57 did the misinformation start after the language restrictions because of hate speech? Do you have a skiff or do you talk openly?
@steve_wk
@steve_wk Год назад
I admire and respect both of these guys so much, but if you've been listening to other such talks recently, this is one boring conversation.
@JaredDudley24
@JaredDudley24 Год назад
GPT has taken over my life my sleep hates GPT! It’s democratizing learning and research!
@Nesamag
@Nesamag Год назад
South Africa loves ChatGPT and users feel a bit left out of the global conversation...
@getyourbone
@getyourbone Год назад
This whole ai thing seems to be a bit of an echo chamber… I am a handy man and scenic builder…how is AI really goin to help me replace a broken ceiling fan? How is ai going to help me build a fake volcano for a corporate event…🤷🏻‍♂️😵‍💫🤔
@chriswayneevans
@chriswayneevans Год назад
AI won’t replace workers like you. Anything requiring human tasks, like in your example, won’t be effected. It’s writers, journalists, software engineers, data entry and other repetitive jobs that are most at risk…initially! But AI will improve over time and businesses want to save money at the end of the day, so AI will expand its scope on the job market in my opinion
@ChrisStewart2
@ChrisStewart2 Год назад
Well, it could copy someone else's volcano design and give it to you. Or it could amalgamate several volcano images into variations which could help
@lolongo
@lolongo Год назад
These guys are just praising each other lol. ask some critical questions! give some honest answers!
@jsfnnyc
@jsfnnyc Год назад
Too much whining about safety when you don’t even know how the models are trained.
@perer005
@perer005 Год назад
That's the point isn't it. You can't sell drugs without showing to the FDA that they are safe.
@cowboy9691
@cowboy9691 Год назад
Reid Hoffman?? No thanks!!!
Далее
OpenAI CEO Sam Altman on the Future of AI
22:57
Просмотров 521 тыс.
Мама приболела😂@kak__oska
00:16
Просмотров 290 тыс.
Italians vs @BayashiTV_  SO CLOSE
00:30
Просмотров 3,5 млн
The moment we stopped understanding AI [AlexNet]
17:38
Просмотров 935 тыс.
Did AI Just End Music? Ft. Rick Beato
25:46
Просмотров 701 тыс.
Мама приболела😂@kak__oska
00:16
Просмотров 290 тыс.