Is AI Progress Stuck? | Jennifer Golbeck | TED





Will progress in artificial intelligence continue to accelerate, or have we already hit a plateau? Computer scientist Jennifer Golbeck interrogates some of the most high-profile claims about the promises and pitfalls of AI, cutting through the hype to clarify what’s worth getting excited about — and what isn’t. (Recorded at TEDxMidAtlantic on June 8, 2024)

If you love watching TED Talks like this one, become a TED Member to support our mission of spreading ideas: https://ted.com/membership

Follow TED!
X: https://twitter.com/TEDTalks
Instagram: https://www.instagram.com/ted
Facebook: https://facebook.com/TED
LinkedIn: https://www.linkedin.com/company/ted-conferences
TikTok: https://www.tiktok.com/@tedtoks

The TED Talks channel features talks, performances and original series from the world’s leading thinkers and doers. Subscribe to our channel for videos on Technology, Entertainment and Design — plus science, business, global issues, the arts and more. Visit https://TED.com to get our entire library of TED Talks, transcripts, translations, personalized talk recommendations and more.

Watch more: https://go.ted.com/jennifergolbeck

TED’s videos may be used for non-commercial purposes under a Creative Commons License, Attribution–Non Commercial–No Derivatives (or the CC BY – NC – ND 4.0 International) and in accordance with our TED Talks Usage Policy: https://www.ted.com/about/our-organization/our-policies-terms/ted-talks-usage-policy. For more information on using TED for commercial purposes (e.g. employee learning, in a film or online course), please submit a Media Request at https://media-requests.ted.com

#TED #TEDTalks

source

32 Replies to “Is AI Progress Stuck? | Jennifer Golbeck | TED”

Well, great stuff overall but i dont follow the keeping 2 developers argument though – they are typically cost centers not revenue centers like sales – so why would the company not want to cut the costs to have more profit? The typical explanation that they can just do / produce more useful output but that depends on whether the company can utilize or sell that output, etc – if they can they will keep the second dev but if there are clear limits on how much dev work they need to have done, they will gladly cut their costs. What am i missing? 🙂

She has a few good points, but I still think she is wrong. I bet if you told the 80% of the population that worked in agricultural before the industrial revolution that in a couple of 100 years only 1% of the population would grow 10x the food that they grew now they would have though you mad.

As intelligent people point out, AI is only going in one direction, which is that it's getting better and more powerful at a rapid rate. This video is only two months old and already it has dated badly. The plateau claim has already been shown to be false. This video is something of an own-goal -a rather stupid mistake. In fact, since this video was made we now have Deep Seek, which came as a surprise and a shock to the AI world, and tonight Feb 17th 2025 we expect Grok 3 to be released -very possibly the most power AI….so far. As I've said, the video has already dated badly.

Considering that I hear the claim that the larger long-term risks detract from the smaller near-term ones in virtually every debate, it's pretty difficult to take seriously. On the contrary, both have helped raise widespread awareness for the other rapidly.

She is totally right.

The main idea behind current AI is learning from an input of data and – based on the learning outcome – calculating probabilities of what is going to happen next. Adding more data to learn from and more computational powers will only decrease the probability of hallucination, but it is technically and physically impossible that the probabilities will be zero. AI hallucination is inevitable and anyone who says otherwise doesn't fully understand what AI is!!

You need infinite learning data and infinite computational power to to get 0% hallucination probability, we will never have either.

And this only applies to current AI, maybe in the future a new concept will emerge that will not depend on learning and calculating probabilities. Thats a different story.

Jennifer: AI is not so dangerous!
Military researcher developing autonomous lethal weapons: Hold my beer..
What an astonishing display of ignorance, lack of domain knowledge and analytical thinking.

For your information. It was in Ukraine on 9 February 2024 that the world's first historical painting with 4 photos was presented in London and Athens during the war, created by the AI algorithm XFutuRestyle based on GPT-4

Leave a Reply