Sam Lehman: What the Reinforcement Learning Renaissance Means for Decentralized AI

Sam Lehman: What the Reinforcement Learning Renaissance Means for Decentralized AI

Join Tommy Shaughnessy from Delphi Ventures as he hosts Sam Lehman, Principal at Symbolic Capital and AI researcher, for a deep dive into the Reinforcement Learning (RL) renaissance and its implications for decentralized AI. Sam recently authored a widely discussed post, "The World's RL Gym", exploring the evolution of AI scaling and the exciting potential of decentralized networks for training next-generation models.

The World’s RL Gym: https://www.symbolic.capital/writing/the-worlds-rl-gym



🎯 Key Highlights


The three phases of AI scaling: Pre-training, Inference Time Compute, and the RL Renaissance.

How DeepMind's novel RL approach (using GRPO) created powerful reasoning models with minimal human data.

Understanding "reasoning traces" and how models learn to "think" longer and more effectively.

The potential downsides of human preference data potentially inhibiting model creativity, drawing parallels to AlphaGo.

Exploring the "World's RL Gym" concept: Decentralizing RL through open environments, diverse tasks, and verified data.

Why open, collaborative RL environments might outperform closed-source labs in generating diverse AI strategies.

The critical role of high-quality base models for successful RL fine-tuning.

Future AI architectures: Continuous learning and the potential of modular Mixture-of-Experts (MoE) models.

Current landscape: Open-source vs. proprietary AI, the challenge of model lock-in, and the role of crypto networks.

Debunking recent claims that "RL is dead" and understanding its true impact.



💡 Want to stay updated with the latest in crypto & AI? Hit subscribe and the notification bell! 🔔



🧠 Follow the Alpha


Tommy's Twitter: @Shaughnessy119

Sam's Twitter: @SPLehman

Symbolic Capital’s Twitter: @symbolicvc



🔗 Connect with Delphi


🌐 Portal: https://delphidigital.io/

🐦 Twitter: https://twitter.com/delphi_digital

💼 LinkedIn: https://www.linkedin.com/company/delphi-digital



🎧 Listen on


Spotify: https://open.spotify.com/show/62PR1RigLG2YN5Pelq6UY9?si=18ac7ccf36ab4753

Apple Podcasts: https://podcasts.apple.com/us/podcast/the-delphi-podcast/id1438148082

Youtube: https://www.youtube.com/channel/UC9Yy99ZlQIX9-PdG_xHj43Q



Timestamps


00:00 - Introduction: Sam Lehman, Symbolic Capital & "The World's RL Gym"

01:30 - History of AI Scaling: Pre-training Era

03:30 - Phase 2: Inference Time Compute Scaling

09:30 - Phase 3: The RL Renaissance & DeepMind Moment

14:30 - How DeepMind Trained R1 without Human Preferences

16:30 - AlphaGo Analogy: Human Data Inhibiting Creativity?

20:30 - Generalizability of RL Training: How Far Does It Go?

22:30 - The "Aha Moment": Models Learning to Think Longer

25:30 - Concept: Decentralized RL & The World's Gym

31:30 - Why Decentralize RL? Open Collaboration vs. Closed Labs

35:00 - Understanding Reasoning Traces

39:00 - Current Decentralized RL Projects (Prime Intellect, General Reasoning)

41:30 - Future Architectures: Continuous Improvement & Modular Models

46:30 - Open Source vs. Proprietary AI: Landscape & Challenges

50:30 - The Lock-In Problem with Foundational Models

52:30 - Is AGI Here? Experiences with GPT-4o

56:30 - Investment Focus in Decentralized AI

59:00 - Modular MoE Models & Jensen's HDEE Paper

1:03:00 - Debunking "RL is Dead" Claims

1:06:00 - Importance of Performant Base Models for RL



Disclaimer


This podcast is strictly informational and educational and is not investment advice or a solicitation to buy or sell any tokens or securities or to make any financial decisions. Do not trade or invest in any project, tokens, or securities based upon this podcast episode. The host and members at Delphi Ventures may personally own tokens or art that are mentioned on the podcast. Our current show features paid sponsorships which may be featured at the start, middle, and/or the end of the episode. These sponsorships are for informational purposes only and are not a solicitation to use any product, service or token.

Avsnitt(468)

Katherine Wu: Breaking Down The EOS and Block One Settlement

Katherine Wu: Breaking Down The EOS and Block One Settlement

Host Tom Shaughnessy of Delphi Digital (DelphiDigital.io) is joined by Notation Capital's Katherine Wu for an awesome discussion on EOS and the Block One settlement. Katherine notated the entire massive document so we walk through what it means, the implications and so much more. This was a fun and insightful episode, one of my favorites.   To access the insights package of Delphi's leading crypto research, visit DelphiDigital.io on your device and sign up using coupon code CHAINREACTION Follow Tom on Twitter @Shaughnessy119 Follow Katherine on Twitter @katherineykwu   Resources: Katherine's Notations on the order and settlement letter Thank Katherine for her work and send a donation! Brians Tweet Disclosures: This podcast is strictly informational and educational and is not investment advice or a solicitation to buy or sell any tokens or securities or to make any financial decisions. Do not trade or invest in any project, tokens, or securities based upon this podcast episode. The host may personally own tokens that are mentioned on the podcast. Tom owns tokens in ETH, BTC, XTZ, LEO, DCR and STX.   - Advertisers: To advertise on this podcast, email Tom@DelphiDigital.io Potential Guests: If you're interested in appearing on the podcast, email Tom@DelphiDigital.io

11 Okt 201937min

Educational Series Part 2 - Scalability with Dan Zuller

Educational Series Part 2 - Scalability with Dan Zuller

Host Tom Shaughnessy of Delphi Digital (DelphiDigital.io) is joined by Vision Hill Group's Dan Zuller for the second part of a multi-part, short and sweet educational series simplifying complex crypto topics. The first episode is focused on scaling blockchain networks on layer 1 and layer 2.  To access the insights package of Delphi's leading crypto research, visit DelphiDigital.io on your device and sign up using coupon code CHAINREACTION Follow Tom on Twitter @Shaughnessy119 Follow Dan on Twitter @danzuller   Resources: Disclosures: This podcast is strictly informational and educational and is not investment advice or a solicitation to buy or sell any tokens or securities or to make any financial decisions. Do not trade or invest in any project, tokens, or securities based upon this podcast episode. The host may personally own tokens that are mentioned on the podcast. Tom owns tokens in ETH, BTC, XTZ, LEO, DCR and STX.   - Advertisers: To advertise on this podcast, email Tom@DelphiDigital.io Potential Guests: If you're interested in appearing on the podcast, email Tom@DelphiDigital.io  -

7 Okt 201924min

Harry Sudock and Nick Sandomeno: Down The Bitcoin Mining and Security Rabbit Hole

Harry Sudock and Nick Sandomeno: Down The Bitcoin Mining and Security Rabbit Hole

Host Tom Shaughnessy of Delphi Digital (DelphiDigital.io) is joined by Harry Sudock, a director of strategy at Griid Infrastructure (Bitcoin Mining Firm) and Nick Sandomeno who is a financial planning analyst at Astorino Financial Group.  We dive into the Bitcoin mining industry, costs, and explore the major issues around mining from sustainability to costs and what the future looks like. This was a very exciting discussion, and sheds light on the mining industry which underpins Bitcoin's security.    Follow Tom on Twitter @Shaughnessy119 Follow Harry on Twitter @harry_sudock  Follow Nick on Twitter @NickSandomeno Links Disclosures: This podcast is strictly informational and educational and is not investment advice or a solicitation to buy or sell any tokens or securities or to make any financial decisions. Do not trade or invest in any project, tokens, or securities based upon this podcast episode. The host may personally own tokens that are mentioned on the podcast. Tom owns tokens in ETH, BTC, XTZ, LEO, DCR and STX.   - Advertisers: To advertise on this podcast, email Tom@DelphiDigital.io Potential Guests: If you're interested in appearing on the podcast, email Tom@DelphiDigital.io  -

4 Okt 20191h 3min

Alex Lindgren: Crypto Regulatory Complexities and Consequences

Alex Lindgren: Crypto Regulatory Complexities and Consequences

Host Tom Shaughnessy of Delphi Digital (DelphiDigital.io) Alex Lindgren, a Partner at Lindgren, Lindgren, Oehm & You LLP. On this episode we discuss the legal complexities projects and teams in the space have to navigate and the consequences.  Mr. Lindgren’s practice ranges from software-based start-ups to private investment funds focusing on cryptocurrency trading or technology acquisitions.   Follow Tom on Twitter @Shaughnessy119 Follow Alex on Twitter @Alex_LLOYLaw Links Alex's Website Disclosures: This podcast is strictly informational and educational and is not investment advice or a solicitation to buy or sell any tokens or securities or to make any financial decisions. Do not trade or invest in any project, tokens, or securities based upon this podcast episode. The host may personally own tokens that are mentioned on the podcast. Tom owns tokens in ETH, BTC, XTZ, LEO, DCR and STX.   - Advertisers: To advertise on this podcast, email Tom@DelphiDigital.io Potential Guests: If you're interested in appearing on the podcast, email Tom@DelphiDigital.io  -

30 Sep 201944min

Joey Krug: Predicting Augur’s Future and Pantera’s Investment Thesis

Joey Krug: Predicting Augur’s Future and Pantera’s Investment Thesis

Host Tom Shaughnessy of Delphi Digital (DelphiDigital.io) is joined by Joey Krug, the co-founder of the predictions protocol Augur and the Co-CIO of Pantera, one of the space's largest group of funds.  We discuss the future of Augur ranging from capabilities, upgrades, Augur V2. We also dive into Pantera's investment thesis, which investments pass the smell test, which investments are worth passing on and so much more. Joey shares the inside scoop on Pantera's investment process and thesis.   Follow Tom on Twitter @Shaughnessy119 Follow Joey on Twitter @joeykrug Links Pantera Capital Augur   Disclosures: This podcast is strictly informational and educational and is not investment advice or a solicitation to buy or sell any tokens or securities or to make any financial decisions. Do not trade or invest in any project, tokens, or securities based upon this podcast episode. The host may personally own tokens that are mentioned on the podcast. Tom owns tokens in ETH, BTC, XTZ, LEO, DCR, VRA and STX.   - Advertisers: To advertise on this podcast, email Tom@DelphiDigital.io Potential Guests: If you're interested in appearing on the podcast, email Tom@DelphiDigital.io  -

24 Sep 201951min

Cosmos’ Sunny Aggarwal: Byzantine Battalion Is Hacking Projects For The Greater Good

Cosmos’ Sunny Aggarwal: Byzantine Battalion Is Hacking Projects For The Greater Good

Host Tom Shaughnessy of Delphi Digital (DelphiDigital.io) is joined by Sunny Aggarwal, a researcher and core developer at Tendermint and Cosmos and the founder of the Byzantine Battalion.  We discuss Sunny's latest project: how the Byzantine Battalion is hacking projects for the greater good. The Battalion can hack projects to test their merits, security and so much more.    Follow Tom on Twitter @Shaughnessy119 Follow Sunny on Twitter @sunnya97 Links The Byzantine Battalion Telegram Group   Disclosures: This podcast is strictly informational and educational and is not investment advice or a solicitation to buy or sell any tokens or securities or to make any financial decisions. Do not trade or invest in any project, tokens, or securities based upon this podcast episode. The host may personally own tokens that are mentioned on the podcast. Tom owns tokens in ETH, BTC, XTZ, LEO, DCR, VRA and STX.   - Advertisers: To advertise on this podcast, email Tom@DelphiDigital.io Potential Guests: If you're interested in appearing on the podcast, email Tom@DelphiDigital.io  -

9 Sep 201933min

SKALE’s Jack O’Holleran: Elastic Sidechains Enabling Massive Scale for dApps On Ethereum

SKALE’s Jack O’Holleran: Elastic Sidechains Enabling Massive Scale for dApps On Ethereum

Host Tom Shaughnessy of Delphi Digital (DelphiDigital.io) is joined by Jack O'Holleran, the co-founder and CEO of SKALE. On this episode we dive into how SKALE's elastic sidechains can enable massively scalable decentralized applications. Between sidechains and state channels, SKALE makes interesting tradeoffs to achieve the right balance of scale and decentralization. SKALE's tech can be deployed with only a few lines of code, elastic sidechains are highly configurable and the system eliminates a lot of the complexity in scaling dApps. Whoever said Ethereum and dApps cant scale haven't heard of SKALE.   To access the insights package of Delphi's leading crypto research, visit DelphiDigital.io on your device and sign up using coupon code CHAINREACTION Follow Tom on Twitter @Shaughnessy119 Follow  jack on Twitter @jackoholleran   Disclosures: This podcast is strictly informational and educational and is not investment advice or a solicitation to buy or sell any tokens or securities or to make any financial decisions. Do not trade or invest in any project, tokens, or securities based upon this podcast episode. The host may personally own tokens that are mentioned on the podcast. Tom owns tokens in ETH, BTC, XTZ, LEO, DCR, VRA and STX.   - Advertisers: To advertise on this podcast, email Tom@DelphiDigital.io Potential Guests: If you're interested in appearing on the podcast, email Tom@DelphiDigital.io  -

3 Sep 201946min

Three Arrows Capital’s Kyle Davies: A Multi-Strategy Fund Seeking Alpha and The LEO Bull Case

Three Arrows Capital’s Kyle Davies: A Multi-Strategy Fund Seeking Alpha and The LEO Bull Case

Host Tom Shaughnessy of Delphi Digital (DelphiDigital.io) is joined by Kyle Davies, the co-founder of Three Arrows Capital. This was an excellent discussion into Kyle's storied finance background and how he is using that to drive alpha in Crypto. Three Arrows Capital goes way beyond long and short positions so it was great to hear all of the ways the fund is driving returns across an array of methods. We also get into the case for LEO, Bitfinex's native token. To access the insights package of Delphi's leading crypto research, visit DelphiDigital.io on your device and sign up using coupon code CHAINREACTION Follow Tom on Twitter @Shaughnessy119 Follow  Kyle on Twitter @kyled116  Follow Su Zhu on Twitter @zhusu  Resources: Polkadot Promise and Problems Three Arrows Capital Website   Disclosures: This podcast is strictly informational and educational and is not investment advice or a solicitation to buy or sell any tokens or securities or to make any financial decisions. Do not trade or invest in any project, tokens, or securities based upon this podcast episode. The host may personally own tokens that are mentioned on the podcast. Tom owns tokens in ETH, BTC, XTZ, LEO, DCR, VRA and STX.   - Advertisers: To advertise on this podcast, email Tom@DelphiDigital.io Potential Guests: If you're interested in appearing on the podcast, email Tom@DelphiDigital.io  -

20 Aug 201947min

Populärt inom Teknik

uppgang-och-fall
elbilsveckan
rss-racevecka
market-makers
natets-morka-sida
skogsforum-podcast
rss-elektrikerpodden
bli-saker-podden
rss-uppgang-och-fall
rss-laddstationen-med-elbilen-i-sverige
rss-veckans-ai
mediepodden
developers-mer-an-bara-kod
solcellskollens-podcast
rss-technokratin
har-vi-akt-till-mars-an
rss-fabriken-2
bilar-med-sladd
hej-bruksbil
rss-kack-tech-podcast