The Inside Story of ChatGPT’s Astonishing Potential | Greg Brockman | TED
TED0 HN recommenders · 2023–2023
1.9M views · talk · news · Uploaded 2023-04-20
OpenAI cofounder Greg Brockman explains the design principles behind ChatGPT and demos unreleased plug-ins at TED 2023. In the follow-up, TED head Chris Anderson presses him on the timeline of development and the risks of releasing such a tool.
HN commenters cite it as a primary source for Brockman's views, such as moving fast to reduce risk, and for capability anecdotes like GPT-4's arithmetic.
15:01What Happened to Horses Is Happening to Us
CGP Grey
27:14Transformers, the tech behind LLMs | Deep Learning Chapter 5
3Blue1Brown
3:31:24Deep Dive into LLMs like ChatGPT
Andrej Karpathy
1:56:20Let's build GPT: from scratch, in code, spelled out.
Andrej Karpathy
1:31:13A Hackers' Guide to Language Models
Jeremy Howard
10:36Why Would AI Want to do Bad Things? Instrumental Convergence
Robert Miles AI Safety
59:48[1hr Talk] Intro to Large Language Models
Andrej Karpathy
1:15:11Veritasium: What Everyone Gets Wrong About AI and Learning – Derek Muller Explains
Perimeter Institute for Theoretical Physics
4 comments from Hacker News
Presumably ChatGPT is better at writing database queries or answering questions about a large Csv file than somebody who doesn’t know how to program. This was the subject of a PR talk last week be an openai cofounder - [VIDEO] - my guess is something like that is what inspired this article.
Based on my understanding from the recent TED talk [0] they mention that GPT-4 can for example add 40 digit numbers perfectly, but makes mistakes on adding a 40 digit and a 35 digit number. It hasn’t worked out a generic understanding of arithmetic but several smaller case specific ones which may not always be correct. [0] [VIDEO]
What the author is describing is simply employers resetting their expectations of employable skills. I've personally seen this happen about 3-4 times over the course of my 30 year career as a software developer. Frankly, the author's tone strikes me as alarmist (and a bit whiny). Yes -- it's uncomfortable. But it's also necessary. Learn the tools well enough to be comfortable with them. And get ready for more change. ChatGPT is already iterating: [VIDEO] and it's not the only General AI tool widely available.
> is the “more safety” camp saying to slow down AI development? This is an EXCELLENT question and the answer is very very nuanced. Warning: the definitions of the terms I'm about to use are actively evolving in current discourse. The following taxonomy reflects my own views, specifically with the inclusion of faction 4: --- There are basically four competing factions. (1) The "normal" faction which includes Sam Altman and which basically says that we can deal with x-risk later and right now let's make money and continue life as normal. (2) The "decel" faction (short for "decelerate"), which includes most old-school AI safety folks such as Ilya, Helen, and Tasha. Sometimes you see these people with a "pause button" emoji or "stop sign" emoji in their Twitter name. (3) The "e/acc" faction (short for "effective accelerationists"). This faction is a mix of fanatical techno-utopians (like Yann LeCun and Andrew Ng), mixed with a bunch of Twitter people who post macho memes and have a "lol let's watch the world burn" sort of attitude. Those people are in my view very similar to the young people from 4chan who voted for Trump over Hillary in 2016 because they thought that a Trump presidency would be hilarious. (4) The newest faction doesn't even have a name. I've only heard it articulated by Greg Brockman so let's call it Brockism. This faction is very new and it actually has me reconsidering my own beliefs. Brockists believe that the safest way to reduce x-risk is to move as fast as