- 26
- 1 358
The Retort AI Podcast
United States
เข้าร่วมเมื่อ 8 ก.ย. 2023
Distilling the major events and challenges in the world of artificial intelligence and machine learning, from Thomas Krendl Gilbert and Nathan Lambert.
retortai.com/
retortai.com/
The Retort's biggest AI stories of 2024 | Retort Episode 35
We're back! Tom and Nate catch up after the Thanksgiving holiday. Our main question was -- what were the biggest AI stories of the year? We touch on the core themes of the show: infrastructure, AI realities, and and antitrust. The power buildout to scale out AI is going to have very real long-term impacts.
Some links this week:
* Ben Thompson's, The End of the Beginning: stratechery.com/2020/the-end-of-the-beginning/
* Miles Brundage's Substack: milesbrundage.substack.com/p/why-im-leaving-openai-and-what-im
* Stochastic Parrots paper: dl.acm.org/doi/10.1145/3442188.3445922
Thanks for listening!
Get The Retort (retortai.com/)…
… on TH-cam: www.youtube.com/@TheRetortAIPodcast
… on Spotify: open.spotify.com/show/0FDjH8ujv7p8ELZGkBvrfv?si=fa17a4d408f245ee
… on Apple Podcasts: podcasts.apple.com/us/podcast/the-retort-ai-podcast/id1706223190
… Follow Interconnects: www.interconnects.ai/
… email us: mail@retortai.com
Some links this week:
* Ben Thompson's, The End of the Beginning: stratechery.com/2020/the-end-of-the-beginning/
* Miles Brundage's Substack: milesbrundage.substack.com/p/why-im-leaving-openai-and-what-im
* Stochastic Parrots paper: dl.acm.org/doi/10.1145/3442188.3445922
Thanks for listening!
Get The Retort (retortai.com/)…
… on TH-cam: www.youtube.com/@TheRetortAIPodcast
… on Spotify: open.spotify.com/show/0FDjH8ujv7p8ELZGkBvrfv?si=fa17a4d408f245ee
… on Apple Podcasts: podcasts.apple.com/us/podcast/the-retort-ai-podcast/id1706223190
… Follow Interconnects: www.interconnects.ai/
… email us: mail@retortai.com
มุมมอง: 117
วีดีโอ
The Nobel AI Albatross
มุมมอง 422 หลายเดือนก่อน
Tom and Nate catch up on the happenings in AI. Of course, we're focused on the biggest awards available to us as esteemed scientists (or something close enough) the Nobel Prizes! What does it mean in the trajectory of AI for Hinton and Hassabis to carry added scientific weight. Honestly, feels like a sinking ship. Some links: * Schmidhuber tweet: x.com/SchmidhuberAI/status/1844022724328394780 *...
Claude Needs a Constitutional Convention | Episode 33
มุมมอง 273 หลายเดือนก่อน
Tom and Nate catch up on recent events (before the OpenAI o1 release) and opportunities in transparency/policy. We recap the legendary scam of Matt from IT department, why disclosing the outcomes of process is not enough, and more. This is a great episode on understanding why the process technology was birthed from is just as important as the outcome! Some links: * Nathan's post on Model Specs ...
Avoiding the AI burnout | Episode 32
มุมมอง 453 หลายเดือนก่อน
Tom and Nate catch up on core themes of AI after a somewhat unintended summer break. We discuss the moral groundings and philosophy of what we're building, our travels, The Anxious Generation, AGI obsessions, an update on AI Ethics vs. AI Safety, and plenty more in between. As always, contact us at mail@retortai.com Some links we mention in the episode: * The Emotional Dog and its Rational Tail...
What we are getting wrong about AI regulation | Episode 31
มุมมอง 284 หลายเดือนก่อน
Tom and Nate catch up on the rapidly evolving (and political) space of AI regulation. We cover CA SB 1047, recent policing of data scraping, presidential appointees, antitrust intention vs. implementation, FLOP thresholds, and everything else touching the future of large ML models. Nate's internet cut out, so this episode ends a little abruptly. Reach out with any questions to mail at retortai....
AI, feedback, and population public health | Episode 30
มุมมอง 345 หลายเดือนก่อน
Tom and Nate revisit one of their old ideas AI through the lens of public health infrastructure, and especially alignment. Sorry about Tom's glitchy audio, I figured it out after the fact that he was talking into the microphone at the wrong angle. Regardless, here are some links for this week. Links: - Data foundry for AI scale.com/blog/scale-ai-series-f - Information piece on Scale AI ($) www....
Apple sends a memo to the AGI faithful | Episode 29
มุมมอง 365 หลายเดือนก่อน
Tom and Nate caught up last week (sorry for the editing delay) on the big two views of the AI future: Apple Intelligence and Situational Awareness (Nationalistic AI doom prevention). One of our best episodes, here are the links: * The Kekulé Problem en.wikipedia.org/wiki/The_Kekulé_Problem * Truth and Method en.wikipedia.org/wiki/Truth_and_Method * Situational Awareness situational-awareness.ai...
Murky waters in AI policy | Episode 28
มุมมอง 256 หลายเดือนก่อน
Tom and Nate catch up on many AI policy happenings recently. California's "anti open source" 1047 bill, the senate AI roadmap, Google's search snaifu, OpenAI's normal nonsense, and reader feedback! A bit of a mailbag. Enjoy. 00:00 Murky waters in AI policy 00:33 The Senate AI Roadmap 05:14 The Executive Branch Takes the Lead 08:33 California's Senate AI Bill 22:22 OpenAI's Two Audiences 28:53 T...
ChatGPT talks: diamond of the season or quite the scandal? | Episode 27
มุมมอง 2026 หลายเดือนก่อน
Tom and Nate discuss two major OpenAI happenings in the last week. The popular one, the chat assistant, and what it reveals about OpenAI's worldview. We pair this with discussion of OpenAI's new Model Spec, which details their RLHF goals: cdn.openai.com/spec/model-spec-2024-05-08.html This is a monumental week for AI. The product transition is completed, we can't just be researchers anymore. 00...
Three pillars of AI power | Episode 26
มุมมอง 317 หลายเดือนก่อน
Tom and Nate discuss the shifting power landscape in AI. They try to discern what is special about Silicon Valley's grasp on the ecosystem and what other types of power (e.g. those in New York and Washington DC) will do to mobilize their influence. Here's the one Tweet we referenced on the FAccT community: KLdivergence/status/1653843497932267520 00:00: Introduction and Cryptozoologi...
Llama 3: Can't Compete with a Capuchin | The Retort Episode 25
มุมมอง 527 หลายเดือนก่อน
Tom and Nate cover the state of the industry after Llama 3. Is Zuck the best storyteller in AI? Is he the best CEO? Are CEOs doing anything other than buying compute? We cover what it means to be successful at the highest level this week. Links: Dwarkesh interview with Zuck www.dwarkeshpatel.com/p/mark-zuckerberg Capuchin monkey en.wikipedia.org/wiki/Capuchin_monkey 00:00 Introductions & advice...
Into the AI Trough of Disillusionment | Episode 24
มุมมอง 878 หลายเดือนก่อน
Tom and Nate catch up after a few weeks off the pod. We discuss what it means for the pace and size of open models to get bigger and bigger. In some ways, this disillusionment is a great way to zoom our into the big picture. These models are coming. These models are getting cheaper. We need to think about risks and infrastructure more than open vs. closed. 00:00 Introduction 01:16 Recent develo...
AI's Eras Tour | Retort Episode 23
มุมมอง 158 หลายเดือนก่อน
Tom and Nate catch up on the ridiculous of Nvidia GTC, the lack of trust in AI, and some important taxonomies and politics around governing AI. Safety institutes, reward model benchmarks, Nathan's bad joke delivery, and all the normal good stuff in this episode! The Taylor moment: DrJimFan/status/1769817948930072930 00:00 Intros and discussion on NVIDIA's influence in AI and the Bay...
Claude 3: Arms race or LLMs stuck in the status quo
มุมมอง 1359 หลายเดือนก่อน
Tom and Nate sit down to discuss Claude 3 and some updates on what it means to be open. Not surprisingly, we get into debating some different views. We cover Dune 2's impact on AI and have a brief giveaway at the end. Cheers! Chapters & transcript partially created with github.com/FanaHOVA/smol-podcaster. More at retortai.com. Contact us at mail at domain. 00:00 Introduction and the spirit of o...
Episode 21: Model release therapy session #1
มุมมอง 429 หลายเดือนก่อน
This week Tom and Nate cover all the big topics from the big picture lens. Sora, Gemini 1.5's context length, Gemini's bias backlash, Gemma open models, it was a busy week in AI. We come to the conclusion that we can no longer trust a lot of these big companies to do much. We are the gladiators playing the crowd of AI. This was a great one, I'm proud on one of Tom's all time best jokes. Thanks ...
Episode 20: Waymo vs. the time honored human experiences, vandalism and defacement
มุมมอง 349 หลายเดือนก่อน
Episode 20: Waymo vs. the time honored human experiences, vandalism and defacement
Episode 18: How to OLMost find a Waifu
มุมมอง 3510 หลายเดือนก่อน
Episode 18: How to OLMost find a Waifu
Tom's Story: to get through grad school, become a sperm whale
มุมมอง 1710 หลายเดือนก่อน
Tom's Story: to get through grad school, become a sperm whale
Non-profits need to be businesses too
มุมมอง 2310 หลายเดือนก่อน
Non-profits need to be businesses too
Episode 14: AI is literally the culture war, figuratively speaking
มุมมอง 14311 หลายเดือนก่อน
Episode 14: AI is literally the culture war, figuratively speaking
Episode 13: What I wish someone had told me
มุมมอง 4111 หลายเดือนก่อน
Episode 13: What I wish someone had told me
Ep. 12: Everyone wants fair benchmarks, but nobody wants to lift heavy ass weights
มุมมอง 26ปีที่แล้ว
Ep. 12: Everyone wants fair benchmarks, but nobody wants to lift heavy ass weights
Ep. 11: Cybernetics, Feedback, and Reinventionism in CS
มุมมอง 60ปีที่แล้ว
Ep. 11: Cybernetics, Feedback, and Reinventionism in CS
E3: As Above, So Below: Alchemy, AI Safety, and RLHF
มุมมอง 35ปีที่แล้ว
E3: As Above, So Below: Alchemy, AI Safety, and RLHF
long hair tom cold opening with the naked truth
Maybe the AI singularity is just the merging of basic sciences. On a serious note, it says a lot about the post-GPT days, where rapid experimentation and validation tests reduced the barrier of core domain expertise to contribute to the basic sciences. Maybe that's the bright side and will encourage more young students to cross-pollinate and contribute in different fields.
The world needs more constitutional conventions overall. I think the french revolution happened thanks to the printing press, the same is about to happen with AI. The first country to include AI wins
the hey its doug opener
followed by soylent (chocolate)
the appalachia episode (good)
19:21 the "_____" community vs the existential risk community? ...Fat?
FAccT
@@NL3360 Is this shorthand for the "AI Ethics" folks (cf "AI Safety" folks) (I see it's ACM FAccT)
15:00 I remember these "backroom discussions" happening at some conferences and thinking it was so cheesy, annoying, and egotistical 😒
Thomas intrigues of the Royal Court arc (DC)
Cryptozoologist 📈📈📈
45:00 If this is the age of bureaucrat administrators, who are the insurgent Ming dynasty eunuchs
30:35 the audience is huge 😭 All my mid-20s SF PM friends are Calcanis-pilled, even the ladies
AI 2024: Don't Cook Your Books
8:45 The BigCos slugging it out doesn't phase the hardened veterans of the west-coast scooter micromobility wars of early ZIRP era (Lime, Bird, Spin, Bolt, Lyft, ...)
49:08 Lungfish Tom immediately exits the trough
32:00 tom cooks
It's the Evergreen (company) Evergiven (ship)
I think you guys need to improve your audio. The volume levels between you need to be equalized. I'm not sure how to do other improvements I'm not an audio guy. It kind of sounds like the signal to noise isn't very good, like you're muffled. As a super easy and quick solution you could try running things through a spoken content audio enhancer. I know adobe has one. I think NPR has some guides to how to produce spoken word content well.
Yeah working on it. We have some solution, but even I can see Toms levels low during recording so if its low then AI reconstruction will only do so much. Sorry for the issue!
I thought they were talking about ARM architecture.😢
AI U-Boat = Public data poisoning? 🤔
Seeing it all in 4k … nice
this was great, elitists NYC undertone and all :)
excellent opening joke
Man they sure love to crash. I believe theyve hit three of our trucks already. Im apart of the company who delivers and moves the cars around for waymo. Also we do cruises. We have 300 to move out of san fran. Insane.
Naming is HARD, but not impossible... No-promo to plug, but I just wanted to chime in on the WIP name game. As far as audience/clique names go, Tortoises is solid. (It even comes with an emoji. 🐢 Nice!) However, I would suggest resisting the urge to shorten with (seemingly innocent) nicknames. Tort, obviously, has a bad association in English with a whole category of law dedicate to torts. In modern French "tort" translates to "wrong, injustice, crime" and "torter" is also not great since it means "torture" en français. 🫣 With all the French (ML/AI) connections... I'd skip those nicknames. Sure... the etymology comes back round to retort by way of some Latin, but nuance is so often OBLITERATED in translation. But... if edgy, tortured, badassery = the vibes you seek to cultivate, then I guess, Torter away... Alchemists would be a logical choice. Alchemy one of Tom's favorite touch points and the Retort's logo is basically alchemical paraphernalia. As a clique name Alchemists isn't too bad... as long as it doesn't get shortened to "alchies". 🥴
Dang. Yeah we should discuss alchemists on the next pod.
Gay's "farewell" letter and her responses to Congress constitute an excellent example of oxymoron, which I suppose in a way informs the veracity of the accusation of plagiarism. But leaving that fact aside, I am more interested in the propensity to sophistry of our current era, far from raising questions in the spirit of seeking foundations, or first principles, our current discourse seeks to distract from precisely raising those questions and seeking their answers honestly. In the light of the latter, an alarm arises in my mind when the word ethics is used so lightly, when we talk about ethics, what are we talking about? what is it based on and why would it be valid? for example, why for some it is a fundamental right to kill an unborn child, but a capital crime to confuse a black with a gorilla? On the other hand, what is the relationship between these ethics and scientific research, its transformation into technology and the pace at which this technology is introduced into society? By way of example, why is it not considered ethical to introduce a vaccine into society without first going through a long trial period (let's not talk about COVID), but it is considered perfectly ethical to introduce technology at an ever-increasing pace without knowing the changes that this will cause in society?
Tom here. Agree that we are missing a substantive understanding of ethics today in AI, and especially with the vaccine vs. disruptive tech analogy. I think the latter will have to change, it might require even more major public fiascos before that happens. This sort of thing doesn't change until it has to.
@@TheRetortAIPodcast Unfortunately you are probably right. The worst thing is that when these fiascos happen (and I hope they don't), a lot of damage will already be done (hopefully at least not irreversible), and this situation will probably end up favoring the strongest companies in the field (e.g. Microsoft), while hurting the rest of the companies, and science in general. But the root of all this lies in the ethics underpinning the functioning of our society (not the nice words we can write in some book, but the motivation behind our individual and organized, private and public behaviors).