No Priors Ep. 65 | With Scale AI CEO Alexandr Wang
Вставка
- Опубліковано 20 чер 2024
- Alexandr Wang was 19 when he realized that gathering data will be crucial as AI becomes more prevalent, so he dropped out of MIT and started Scale AI. This week on No Priors, Alexandr joins Sarah and Elad to discuss how Scale is providing infrastructure and building a robust data foundry that is crucial to the future of AI. While the company started working with autonomous vehicles, they’ve expanded by partnering with research labs and even the U.S. government.
In this episode, they get into the importance of data quality in building trust in AI systems and a possible future where we can build better self-improvement loops, AI in the enterprise, and where human and AI intelligence will work together to produce better outcomes.
Sign up for new podcasts every week. Email feedback to show@no-priors.com
Follow us on Twitter: @NoPriorsPod | @Saranormous | @EladGil | @alexandr_wang
0:00 Introduction
3:01 Data infrastructure for autonomous vehicles
5:51 Data abundance and organization
12:06 Data quality and collection
15:34 The role of human expertise
20:18 Building trust in AI systems
23:28 Evaluating AI models
29:59 AI and government contracts
32:21 Multi-modality and scaling challenges - Наука та технологія
Refreshing to listen to a meaningful discussion focusing on the data the AI models are trained and refined on.
I've been theorizing with friends that post-scarcity, post-AGI societies will be more about deeper evolutionary needs and their associated problems, while AI will handle everything else. This CEO's idea that the model's ability to "optimize for amorphous goals over exceedingly long time horizions" is a fundamental limitation seems in line with my evolution theory. We can't re-create the exact pressures that developed human intelligence (i.e. evolution over millenia), so it makes sense that we will never create something artificial that captures all of this context.
AI will do a shit ton, but not *everything*.
It's nice to see so much transparency. I think this would be a good video for the Remotasks community to see. I'm thankful I saw this video. Thank you!
recommend something more please
C00L
Alex is clearly not very bright (or maybe just too young?)
Elad clearly understands this, but cannot say it out loud
I respect Alex for winning, but the founding story of Scale AI is total BS in this video. He was a freshman at MIT in his first semester, he wasn’t working on AI yet. He got into YC with an app like Zocdoc then pivoted into Scale. This whole ‘wow you had so much foresight’ is a load of BS
He was building smth like Opentable but for clinics. He scraped data from websites and this experience inspired him to build APIs for human-in-the-loop services like evaluating a website, extracting info from an article, etc. Eventually the service expanded to data annotation and the rest is history.
You can still have foresight for a new direction. How does that not make sense? You start down one path, realize the future is going down another path and you pivot. That is quite literally foresight.
How many of us get it right on the first idea though?
Nothing wrong with pivoting and changing direction until you get it right, my issue is with how he is telling the story. The way he’s presenting it is very different from the messy reality. It’s ok to admit you were just making it up as you went along
well this was a waste of time..
openai cofounder already said ai will automate his job away in 5 years. so good luck to you if you think humans + ai will beat ai alone
if openai said it, it must be true then. thanks for the heads up.
:)
Bet they smashed when he was in her house
WATT NO WAY
I THOUGHT HE WAS A GEY 😒
He is virgin male sacrifice
WATT NO WAY
I THOUGHT HE WAS A GEY
Just let him be a male virgin sacrifice
No gf
I take that one ty