6 ms·
OpenAI CEO Sam Altman on Lex Fridman Podcast [video]
- tpmx 4y agoThe victimhood claim didn't take long. Sure, that's Lex's editing, but.. Anyway, then read: https://news.ycombinator.com/item?id=34471720 https://news.ycombinator.com/item?id=34471720 > Ask HN: How did Sam Altman fail upward so well? > 245 points by VirusNewbie 62 days ago - 119 comments
- m_ke 4y agoI'm glad Lex chose to spend half of it discussing Jordan Peterson and Elon Musk. Also as great and useful as the LLMs are, it's still a bit of a leap to claim it's getting us that much closer to AGI. GPT does a great job compressing and distilling the knowledge found in text on the internet but it still doesn't have the ability to do real reasoning. As a simple example I recommend trying to have it multiply two large numbers, it does a decent job guestimating but obviously can't do the recursive steps required to generate a correct answer. EDIT: Also people worried about AGI implications should slow down and worry what humans and society will do with tools that can automate a large portion of our jobs and accrue all of the gains of that to a few well capitalized players.
- uxcolumbo 4y agoHas Lex started to be more critical of Musk or is he still behaving quasi sycophantic?
- sebzim4500 4y ago>As a simple example I recommend trying to have it multiply two large numbers, it does a decent job guestimating but obviously can't do the recursive steps required to generate a correct answer. Is it worse than the average human at mental arithmetic? I haven't done detailed benchmarks, but for my usecase it only gets maths wrong once it gets way past the range that I could do myself. Personally, I find the fact that it can't say "I don't know" to be much more offputting than the dodgy maths.
- ftxbro 4y ago> Personally, I find the fact that it can't say "I don't know" to be much more offputting This is a problem with their PPO "finishing school" not with their base model. Check Figure 8 in https://arxiv.org/pdf/2303.08774.pdf https://arxiv.org/pdf/2303.08774.pdf its base model has an uncannily well calibrated confidence in the correctness of its decisions. Basically, their PPO is for training it to talk to normies, and normies don't like uncertainty. Reinforcement learning using Proximal Policy Optimization (PPO): They optimize a policy against the reward model using reinforcement learning. For each new prompt sampled from the dataset, the policy generates an output. The reward model calculates a reward for the output, and the reward is used to update the policy using the PPO algorithm.
- sebzim4500 4y agoTrue, but even with the base model it is hard to go from logits to a confidence level for freeform answers. For multiple choice questions it is easy, since you just look at the probabilities for each of the allowable tokens in the answer (which is presumably what they did in that benchmark) but I have no idea how to do it more generally.
- r0uv3n 4y agoIt is absolutely not obvious it can't do the recursive steps to multiply two numbers. In fact, you can just teach it the proper algorithm and ask it to apply it: https://twitter.com/random_walker/status/1600336556425826304?s=20 https://twitter.com/random_walker/status/1600336556425826304...
- m_ke 4y agoThe feedforward nature of transformers means that it can only do a limited number of computational steps. Having it explain an algorithm is a good way to have it use its output as a tape during inference/decoding, but even then it just fudges things and makes mistakes for largish numbers.
- ActorNightly 4y ago>it still doesn't have the ability to do real reasoning. It technically does. Humans solve problems in a recursive way, much like the generative transformer networks generate the next piece of text. The big difference of course is the depth of reasoning. My bet is that the next big effort is going to be reduce the size of the models while retaining performance (perhaps using ML tasked to do so). Then once you have that, from the other side there will be more and more specialized hardware with ASIC to run the transformers. Eventually we will be able to run these things at scale and do on the fly transfer learning, which will be about the same as a human growing reasoning skills as they mature.
- ftxbro 4y agoThis parent post with its EDIT is a perfect example of what I wrote in another comment: Beliefs are changing so fast right now. The term "AGI skeptic" will soon (if not already) mean "I don't trust AGIs in positions of authority or power" rather than "I don't think the technology is capable of matching our level of cognition."
- ChatGTP 4y agoHang on, we’ve also changed the definition of AGI, now this thing is increasingly “marketed” as an AGI, and as someone has used it, I don’t want it anywhere near critical decision making. There is marketing and there’s reality. Don’t conflate the two. There is untold money which has been spent on these systems. They’re going to be pushed into our lives whether we like it or not. That’s not to say there is no value in machine learning, but there also a lot of money to be made from it. Think opioid crisis, you had it prescribed even if it wasn’t in your interest. This is America…
- djaychela 4y ago> it still doesn't have the ability to do real reasoning I don't think that matters. It does significantly more than algorithms which are designed to simply maximise engagement, and they have revolutionised the online world in the last decade. This will be orders of magnitude more powerful. And given what the simple 'maximise engagement' algorithms have done to various aspects of society and the people who live in it, I don't think this will be good given how little attention seems to be paid to these aspects by those who stand to profit from their use.
- madeofpalk 4y agoIt might not matter to whether it's a useful product, but it matters as to whether it is artificial general inteligence or not.
- m_ke 4y agoYes, my post was mostly a response to the first few segments of the interview where Sam talks about how experts in the field didn’t believe in their (openai) focus on AGI. I’ve been working in deep learning for over 10 years now and was always bullish on the technology but as great of a tool as these LLMs might be, I wouldn’t bet on them getting us to true AGI anytime soon.
- gautamdivgi 4y agoI find Peterson to be an insufferable coot. Musk is obviously more dynamic.
- 3327 4y ago[dead]
- ilrwbwrkhv 4y agoI find Lex Fridman to be the most boring podcaster on the planet. And I have heard plenty of drones in the past, but he takes it to the next level.
- dark_mode 4y agoi find cats to be annoying, but i don't go around commenting that on cat videos
- ilrwbwrkhv 4y agoBut you do comment on comments about cat videos.
- appstorelottery 4y agoIt's hard not to resist commenting on comments on comments about cat videos.
- cronix 4y agounless someone's pointing a laser at it and wiggling it around
- purple_ferret 4y agoHe's just a bad interviewer. An example I've noticed: A guest will have essentially addressed a question Lex has planned for later in the interview, but Lex still asks it without any sort of modification/acknowledgment of what the guest has already said, leading to a lot of repetitive talk. He also basically ignores statements that conflict with his view.
- ftxbro 4y agoI suspected something was bad, when Fridman managed to interview Demis Hassabis on the topic of "AI, Superintelligence & the Future of Humanity" for more than two hours in July 2022, and the topic of large language models never came up. I don't know the reason. Someone has suggested it was because it was super secret and Hassabis made a deal with Fridman before the interview to not talk about it but ... idk.
- tiffanyh 4y agoOff topic: did Sam get a face-lift? Or is that just tons of Botox?
- newmac 4y agoI rolled my eyes when I saw your comment, then I watched the video. The lips and cheeks are distinctly botox'd.
- uxcolumbo 4y agoWhy would you get botox at 37? Unless it's medical.
- gen220 4y agoI know lots of people in their late 20s / early 30s who get botox. I personally don't understand it either, but it's definitely a thing.
- ushakov 4y agoSelf-esteem issues
- pengaru 4y agoThere's supposedly an increasingly common practice of starting early, so the wrinkles in those areas don't ever develop in the first place... I imagine it also prevents a more abrupt change later on if you're intending to do it. Going from an expressive wrinkled face to a paralyzed one in your 50s would be much more jarring for your peers, than just lacking expressiveness much of your adult life. It's all a rather weird form of vanity to me.
- capableweb 4y agoTimeline of the Podcast: 4:36 - GPT-4 16:02 - Political bias 23:03 - AI safety 43:43 - Neural network size 47:36 - AGI 1:09:05 - Fear 1:11:14 - Competition 1:13:33 - From non-profit to capped-profit 1:16:54 - Power 1:22:06 - Elon Musk 1:30:32 - Political pressure 1:48:46 - Truth and misinformation 2:01:09 - Microsoft 2:05:09 - SVB bank collapse 2:10:00 - Anthropomorphism 2:14:03 - Future applications 2:17:54 - Advice for young people 2:20:33 - Meaning of life Seems like about the first half (up to the 1 hour mark) is interesting enough to warrant a listen, but echoing what ilrwbwrkhv said, Lex Fridman is kind of boring which makes it hard to even sit through the first half. Feels a bit like he has the personality of a fridge.
- sebzim4500 4y ago> Feels a bit like he has the personality of a fridge. Do other people listen to podcasts based on the personality of the interviewer? I just assume that you listen based on the guest, and the host is just there to ask decent questions, which Lex does quite well.
- capableweb 4y agoUsually I listen to interviews based on the chemistry between the two. If either of them are bland, the entire thing gets thrown off balance. If it's all about the guest, the perfect interview for you would be someone emailing the guest the questions so they can read them themselves, without involving a second person?
- alexchro93 4y agoMost definitely. The interview-style podcast I listen to the most is Odd Lots [1]. I'll listen to any Odd Lots episode regardless of the guest because of the personalities of Tracy and Joe. I also listen to re-runs of Car Talk [2] (which, I guess maybe isn't a podcast?) because of the hosts. 1. https://www.bloomberg.com/oddlots https://www.bloomberg.com/oddlots 2. https://www.npr.org/podcasts/510208/car-talk https://www.npr.org/podcasts/510208/car-talk
- madeofpalk 4y ago
- 0xbadc0de5 4y agoLots of good insights in this podcast.
- bluehorseray 4y ago14:17: "I suspect too much of the processing power is going into using the model as a database instead of using the model as a reasoning engine" https://youtu.be/L_Guz73e6fw?t=857 https://youtu.be/L_Guz73e6fw?t=857
- spprashant 4y agoCan you (or someone) expand more on what he means by this?
- wahnfrieden 4y agoasking and retrieving info instead of iterating through analysis steps with inputs toward some goal
- jimsimmons 4y agoWhen you ask GPT to add it probably tries to remember a similar instance in the training set rather than performing the calculation on the spot
- eternalban 4y agoSam indicated, in context of discussing a "democratic" constitution for AI use, that OpenAI can -not- be a "mere advisor" in the process. Sam also (unwittingly) underscored the alarm of the NYTimes columnist [for me] that wondered at the open admission (which Sam repeated here) that "there is a chance" that AI will kill off humanity, but "the iterative approach" of OpenAI could help with that.
- preommr 4y agoIt's remarkable how unremarkable that interview was. I think a major problem I had is that there was very little new substance with lots of nonsense. One of the biggest criticisms about transitioning from non-profit to for-profit can basically be summed up as they "needed to raise capital but don't worry, we're still ethical". Or the ongoing discussions about safety and bias that have been discussed to death at a very superficial level. Or a question about how much parameter size matters is basically "we just want to do what's best". How about some insightful answers from someone with first-hand knowledge instead of answering like a politician running for office? Lots of weird circle-jerking as well about how he's one of the few people that can maybe make AGI and how powerful that makes him. I feel like the interview is completely detached from the real world where it basically comes down to advancements in technology and data. We saw how fast Stable Diffusion covered ground. Also really feels like it's too sensational to be talking so much about AGI - idk why Altman is acting so mysterious about 'democratizing the process'. Make the models public. Make the research public. And of course there's the typical Fridman nonsense. There's a whole section on 'Elon Musk'. Not even really his history with the company, but about twitter drama. It's distracting how much Fridman worships Elon - like he's made multiple offers to elon to run twitter only to get ignored.
- jimsimmons 4y agoWhy didn’t OpenAI just shut down if being a non profit was unviable? Of course great things have happened because it survived. But survival instinct is as much a part of that transition as anything else
- ramraj07 4y agoWelcome to all of human discourse. Has this been any different from any point in human history? If you’re a scientist, 80% of the time you spend in conferences or reading papers would be digesting stuff you already know, with a few new nuggets of information or communication. You have to acknowledge every human being has an agenda and they’re gonna act in their self interest. Why would Sam give more details than he wants to in this or any meeting given the primarily economical but ostensibly ethical reasons not to reveal it? I still got many many useful pieces of information from this, that was worth the 2 hours of time and enduring all the drivel that goes in between by nature of this being a lex podcast.
- ftxbro 4y agoLex Fridman is one of the podcasters on the planet right now, hosting celebrity guests for long form interviews.
- braindead_in 4y agoInteresting comment about consciousness being the fundamental substrate of the world and we are in a dream or simulation. That is exactly the Nondualist position. Did he say Bramhan? https://youtu.be/L_Guz73e6fw?start=1:08:22 https://youtu.be/L_Guz73e6fw?start=1:08:22
- AlchemistCamp 4y agoI had the exact same question. YT subtitles say he did.
- lukeplato 4y agowish there was a follow up discussion about those ideas instead of a context change. And yes, Altman has tweeted something similar before: > haven’t seen this as a twitter thread, so: what true thing do you believe that few people agree with you on? > absolute equivalence of brahman and atman https://twitter.com/sama/status/1607482407770554373 https://twitter.com/sama/status/1607482407770554373
- O__________O 4y agoAppears interview took place on March 20, 2023, which was 6-days after launch of GPT4: https://youtu.be/L_Guz73e6fw?t=1797 https://youtu.be/L_Guz73e6fw?t=1797
- jimsimmons 4y agoWhy does Lex have so much “now impress me” energy in these podcasts? It’s not even a personality quirk because he seems to modulate it throughout the podcast. Surely he can get rid of it altogether. Rise of Lex is both baffling and utterly expected to me as a AI insider. He used his MIT badge to the fullest extent. Having seen his “lectures” I wouldn’t bet on him cracking an interview for a beginner AI researcher role anywhere
- corbulo 4y agoHe seemingly doesn't have...any depth in his public persona. Very bizarre guy. Feels more obfuscatory than actually being that lacking. His comment sections scream inorganic to me. Always 99% blanket sunshine out the ass praise for how mind blowingly incredible every episode is, never with any specific details.
- ChatGTP 4y agoJust want to remind people, Lex is a human being. So is Sam, keep it constructive ?
- mychael 4y agoThe plastic surgery has a heavy uncanny valley vibe.
- bronzebrain 4y ago[flagged]