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)

Delphi Digital’s Kevin Kelly: The Perfect Storm for Bitcoin

Delphi Digital’s Kevin Kelly: The Perfect Storm for Bitcoin

Host Tom Shaughnessy of Delphi Digital (DelphiDigital.io) is joined by my partner and and a co-founder of Delphi Digital itself, Kevin Kelly. Given Kevin cut his teeth as an equity strategist with Bloomberg, he offers unique and differentiated insights on the institutional and macro argument for Crypto and Bitcoin. Kevin offers his take on the global backdrop, everything from underfunded pensions to global monetary policy and even the eventual potential dethroning of king dollar. We go on to where Kevin believes crypto and the global economy will co-exist in the future and what would happen to crypto if a recession hits.  With all of the endless tech and project news out there everyday, this episode was a breath of fresh air as kevin offers an extremely zoomed out view of where Bitcoin fits into our institutional world and broader global monetary framework  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 Kevin on Twitter @Kevin_Kelly_II    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  -

16 Juli 20191h 2min

Harmony’s Nick White: Mainnet Launch Is The First Step Towards Full Proof-of-Stake and Sharding

Harmony’s Nick White: Mainnet Launch Is The First Step Towards Full Proof-of-Stake and Sharding

Host Tom Shaughnessy of Delphi Digital (DelphiDigital.io) is joined by Nick White, the co-founder of Harmony Protocol Harmony is described as a fast and secure blockchain and it leverages a lot of newer tech such as Proof of Stake, Sharding and cross shard communications, granted not all of this tech is fully launched yet.  We discuss Harmony’s ultimate goal to create an open infrastructure for the world's data, a global scale marketplace platform, competition, Harmony’s IEO on binance and so much more.  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 Nick on Twitter @nickwh8te    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 Juli 201951min

Cadence’s Nelson Chu: The Digital Securitization and Investment Platform for Private Credit

Cadence’s Nelson Chu: The Digital Securitization and Investment Platform for Private Credit

Host Tom Shaughnessy of Delphi Digital (DelphiDigital.io) is joined by Nelson Chu, The Founder and CEO of Cadence. For those new to cadence, the company is a Technology-driven securitization platform connecting retail and institutional investors with short term structured private credit investment opportunities. Cadence is using Blockchain Technology to open up this market to smaller investors, using technology to automate share fractionalization and distribution and simplify 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 Cadence on Twitter @withcadence   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.  - 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  -

1 Juli 201936min

IDEX’s Alex Wearn: A Sleeping Giant’s Plan To Dominate Every Centralized Crypto Exchange

IDEX’s Alex Wearn: A Sleeping Giant’s Plan To Dominate Every Centralized Crypto Exchange

Host Tom Shaughnessy of Delphi Digital (DelphiDigital.io) is joined by Alex Wearn, the Co-Founder & CEO of IDEX.   This is a very interesting episode as we discuss how IDEX, a sleeping giant, grew to become the largest decentralized exchange and smart contract on Ethereum. We discuss competition with other DEXs, centralized exchanges like Binance and Coinbase and hybrid competition.  We also dive into IDEX’s future, discussing its two tokens; one membership token and one staking token, the future of governance on IDEX and some details on its layer-2 solution to increase throughput. Frankly, I had no idea IDEX had this much going on, and Alex shares some extremely insightful details around the entire IDEX roadmap. 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 Alex on Twitter @AlexWearn    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 personally owns tokens in ETH, BTC, LEO, XTZ, DCR and MKR  - 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   Intro music by Chris Zabriskie

27 Juni 20191h 5min

Edgeware’s Dillon Chen: Bringing Smart Contracts to Polkadot and ~$650M In Lockdrop

Edgeware’s Dillon Chen: Bringing Smart Contracts to Polkadot and ~$650M In Lockdrop

Host Tom Shaughnessy of Delphi Digital (DelphiDigital.io) is joined by Dillon Chen of Commonwealth Labs to discuss Edgeware. As a reminder Polkadot is an interoperability platform which is set to launch in a few months, and Edgeware’s goal to bring smart contracts to Polkadot itself.  This would bring smart contract functionality to polkadot, and allow it to compete with Ethereum, Tezos, EOS and other smart contract platforms.  Edgeware is currently running an innovative alternative to a raise, known as a lockdrop. Instead of raising money from investors, Edgeware is allowing holders of Ether to lock their ETH up and receive Edgeware tokens in the future or to signal their intent to back the project, without locking to earn edgeware tokens. This is a really innovative approach to not only garner a large potential stakeholder base which is critical for a project's success, but also to increase the value of Edgeware’s future treasury without raising money from investors. Currently there is 881k ETH locked and 1.2M ETH signaled, or $273M and $372M in value respectively.  Edgeware is a high profile project, and this is hopefully the most in-depth discussion to date on Edgeware with a ton of great color on Polkadot as well.    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 Dillon on Twitter @dillchen   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 personally owns tokens in ETH, BTC, LEO, XTZ, DCR and MKR  - 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   Intro music by Chris Zabriskie

25 Juni 201941min

DARMA Capital’s Andrew Keys and James Slazas: $100M Ether Fund Launches - Largest of 2019

DARMA Capital’s Andrew Keys and James Slazas: $100M Ether Fund Launches - Largest of 2019

Host Tom Shaughnessy of Delphi Digital (DelphiDigital.io) is joined by Andrew Keys and James Slazas of DARMA Capital to discuss DARMA's $100M Ether Fund.   This episode was recorded on the day of Andrew’s announcement that he was was joining DARMA, after an impactful career at ConsenSys. DARMA's flagship fund is a $100M Ether optimized long strategy with $100M under management, which is sizeable for the space. On this episode we go into the strategy, both James and Andrew’s careers, their results, goals and so much more. The brain power between James and Andrew is remarkable. 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 Andrew on Twitter @AndrewDARMACAP  Follow James on Twitter @DARMA_Slazas    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.  - 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   Intro music by Chris Zabriskie

19 Juni 201942min

Blockstack’s Muneeb Ali: Building A Decentralized AWS To Host A Potential Facebook Killer

Blockstack’s Muneeb Ali: Building A Decentralized AWS To Host A Potential Facebook Killer

Host Tom Shaughnessy of Delphi Digital (DelphiDigital.io) is joined by Muneeb Ali the CEO and Cofounder of Blockstack PBC. Blockstack is a high profile project in the space, with the goal of creating a decentralized computing network and app ecosystem. To differentiate this episode vs other podcasts, I wanted to focus on Blockstacks’ end goals - being potentially creating a new decentralized version of Amazon Web Services which hosts basically the worlds business and consumer workloads and apps. Creating this platform could in turn host a viral app on top of it, such as a facebook killer. 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 Muneeb on Twitter @muneeb   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.  - 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   Intro music by Chris Zabriskie

17 Juni 201959min

Decred’s Co-Founder Jake Yocom-Piatt: Governance-First Crypto Aims to Challenge Bitcoin

Decred’s Co-Founder Jake Yocom-Piatt: Governance-First Crypto Aims to Challenge Bitcoin

Host Tom Shaughnessy of Delphi Digital (DelphiDigital.io) is joined by Jake Yocom-Piatt, an original co-founder of the cryptocurrency Decred.  It's incredible to speak with Jake and learn about how he built Decred with its anonymous co-founders, all of their struggles and success and where the crypto is going. We go into Decred’s hybrid consensus mechanism, its on-chain treasury, easy voting on consensus changes and the use of the treasury through the wallet decredition and so much more. While Decred is a smaller crypto $250M market cap, but Jake explains why he believes Decred is not only competitive to BTC but could over take it.  Follow Tom on Twitter @Shaughnessy119 Follow Jake on Twitter @behindtext   Disclosure: Tom Shaughnessy owns tokens in BTC, ETH, DCR, MKR, XTZ and Loom. This podcast is NOT investment advice and is only informational. Do not make investment decisions based upon this podcast. Delphi Digital was not compensated by any party for this podcast episode other than Podbean's advertisers. This content is strictly informational and is not a solicitation to buy or sell any security or token.  - 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  -

8 Juni 20191h 6min

Populärt inom Teknik

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