AI, Explained: Why It’s Different This Time | WSJ Tech News Briefing

There is so much being said about artificial intelligence these days. In order to understand this latest wave of AI, it is important to know how it works and why this moment is different from developments in the past.

WSJ science reporter Eric Niiler joins host Zoe Thomas for the first installment of Tech News Briefing’s special series Artificially Minded. New episodes drop every Monday in April.

0:00 AI basics
2:18 What is machine learning?
4:20 How generative AI like ChatGPT works
6:07 AI in our daily lives
7:45 How people on the street feel about AI
8:38 Why AI is a hot topic right now
9:50 Risks and rules surrounding AI

Tech News Briefing
WSJ’s tech podcast featuring breaking news, scoops and tips on tech innovations and policy debates, plus exclusive interviews with movers and shakers in the industry.

#AI #ArtificialIntelligence #WSJ

[piratepartyftw]

I work in AI. This "expert" is not an expert.

He's fundamentally misunderstanding how these Large Language Models work when he says they "just repeat things." That's not correct at all. Under the hood, they are synthesizing what they've "read" into a model of the world. It's blurry and imperfect, but they do "know things." This is why it can answer questions no one has ever asked before, and generate completely new commentaries on existing ideas or texts.

He also underestimates how fast they are advancing and how big of a deal theyre going to be. These LLMs are very quickly (within 2 years) going to be able to do most anything that an average human can do on a computer. White collar work is going to be utterly transformed. Millions and millions of layoffs are coming.

[wetube]

Plot twist: Zoe Thomas is an AI avatar

[kalasatwater]

Read a beautiful tweet the other day: "A world where humans do the hard labour for minimum wage while AI write poetry and create art is not the future I wanted."

[Waterfox]

We will definitely have self driving cars by 2018!

[AiLatestNews]

Insightful perspective! This video delves into why AI is distinct this time. Engaging and informative content.

[ACK]

But I think that is a program, collecting and converting answers in certain time frames and creates confusions without controls.

[lanediver]

the WSJ guy was wrong about ai vs machine learning..."While artificial intelligence encompasses the idea of a machine that can mimic human intelligence, machine learning does not. Machine learning aims to teach a machine how to perform a specific task and provide accurate results by identifying patterns." -- additionally, an algorithm isn't a program, programs implement algorithms...

[sentientcardboarddumpster]

This is moving fast. I feel like the next 10 years is incomprehensible

[AparnaModou]

Some of these apps/software woudn't be considered an actual AI. Bluewillow for exampe, I'd categorize this as a machine learning software using the diffusion model. So over time, these programs would evolve and improve.

[almerammar]

ok they understad syntax, but can they understand semantics?

[shuk]

The biggest problem of AI will be the crisis of meaning.

What do you do when career is out of sight?

You might think, I hate my job, what are you talking about?

But think about it. Even if you hate your job it fulfills you in some way, whether it's money, social events or personal achievement.

This will not happen in early phases obviously but we will find ourselves more and more useless.

Sorry to sound negative but I hope we will make the best out of it.

[SuperGGnoRE]

I disagree with the wording in the explanation because it implies intent or goal-oriented "thinking". These LLMs are still just statistical models. That being said, they are proving to be very capable tools. But that's all they are: tools. I'd be more worried about other people exploiting these tools than any "artificial sentience".

[JordChase]

I'm thrilled to see a dedicated series like "Artificially Minded" exploring the intricacies of artificial intelligence! With so much buzz around AI these days, it's crucial to understand its inner workings and what sets this current wave apart from previous developments. Having WSJ science reporter Eric Niiler join host Zoe Thomas for this insightful discussion is a fantastic choice. I'm eager to dive into this first installment and gain a deeper understanding of AI's impact on our world. Kudos to Tech News Briefing for tackling this important topic and bringing us valuable insights. Can't wait for more episodes in the series! 🤖🎙📚

[mojtabapeyrovian]

AI can provide significant benefits if an automation tax system is implemented, encouraging the utilization of AI as a learning tool to empower individuals rather than solely relying on it as an automation tool that creates difficulties and chaos in life. Therefore companies that choose to use machines instead of human labour may end up with higher expenses for their products and services. On the other hand, businesses that don't rely on automation can still remain competitive

[SoCalFreelance]

AI Monday? Is that like taco Tuesday? 🌮

[erictustison]

It is misleading to present this guest as an expert in AI: his description of LLMs seems to accurately describe earlier models, but does not reflect recent work, such as "Sparks of Artificial General Intelligence: Early experiments with GPT-4" by Bubeck et al or "The Alignment Problem from a Deep Learning Perspective" by Ngo et al. Specifically, SOTA LLMs seem to emergently *understand the meaning of words and can take actions* ; they are not merely "repeating things" (9:27)

[dolevmazker]

So basically its just gonna change some industries for the better while Eric notes how there never was once a case where AI just made decision on its own or took a decision by its own accord, so its not going to annihilate someone unless someone can program an AI to be commanded to hunt/kill or worse. I think that's where the danger is, with people with bad intentions with AI that have that intelligence to implement it.

That guy interviewed summed it up well. " I think its good, but highly skeptical "

[HaHaBIah]

Tech news briefing, the most reliable part of wsj

[RegularRegs]

Looking forward to watching artificially minded

[gamalchamac]

Neural network is another machine learning technique to make predictions based on data and the technique is inspired by the connections of our neurons in the brain. This Eric NIILER just shows that it is not easy to digest these terms