10 Hours of Listening.
5 Minutes of Reading.

Deep dives into the conversations shaping the future of AI, Robotics & Crypto.

Save hours of your time each week with our podcast aggregator

🔍 Search & Filter
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.

AI Podcasts

December 31, 2025

[State of Evals] LMArena's $100M Vision — Anastasios Angelopoulos, LMArena

Latent Space

AI
Key Takeaways:
  1. The transition from static benchmarks to "Vibe-as-a-Service" means model labs must optimize for human delight rather than just loss curves.
  2. Use Arena’s open-source data releases to fine-tune models on real-world prompt distributions.
  3. In a world of synthetic data and benchmark saturation, human preference is the only remaining scarce resource for validating frontier capabilities.
See full notes
December 31, 2025

[State of Context Engineering] Agentic RAG, Context Rot, MCP, Subagents — Nina Lopatina, Contextual

Latent Space

AI
Key Takeaways:
  1. The transition from Model-Centric to Context-Centric AI. As base models commoditize, the value moves to the proprietary data retrieval and prompt optimization layers.
  2. Implement an instruction-following re-ranker. Use small models to filter retrieval results before they hit the main context window to maintain high precision.
  3. Context is the new moat. Your ability to coordinate sub-agents and manage context rot will determine your product's reliability over the next year.
See full notes
December 31, 2025

[NeurIPS Best Paper] 1000 Layer Networks for Self-Supervised RL — Kevin Wang et al, Princeton

Latent Space

AI
Key Takeaways:
  1. The convergence of RL and self-supervised learning. As the boundary between "learning to see" and "learning to act" blurs, the winning agents will be those that treat the world as a giant classification problem.
  2. Prioritize depth over width. When building action-oriented models, increase layer count while maintaining residual paths to maximize intelligence per parameter.
  3. The "Scaling Laws" have arrived for RL. Expect a new class of robotics and agents that learn from raw interaction data rather than human-crafted reward functions.
See full notes
December 31, 2025

[State of AI Papers 2025] Fixing Research with Social Signals, OCR & Implementation — Team AlphaXiv

Latent Space

AI
Key Takeaways:
  1. The Age of Scaling is hitting a wall, leading to a migration toward reasoning and recursive models like TRM that win on efficiency.
  2. Filter your research feed by implementation ease rather than just citation count to accelerate your development cycle.
  3. In a world of AI-generated paper slop, the ability to quickly spin up a sandbox and verify code is the only sustainable competitive advantage for AI labs.
See full notes
December 31, 2025

[State of MechInterp] SAEs in Production, Circuit Tracing, AI4Science, "Pragmatic" Interp — Goodfire

Latent Space

AI
Key Takeaways:
  1. The transition from Black Box to Glass Box AI. Trust is the next moat, and interpretability is the tool to build it.
  2. Use feature probing for high-stakes monitoring. It is more effective and cheaper than using LLMs as judges for tasks like PII scrubbing.
  3. Understanding model internals is no longer just a safety research project. It is a production requirement for any builder deploying AI in regulated or high-stakes environments over the next 12 months.
See full notes
December 31, 2025

[State of Code Evals] After SWE-bench, Code Clash & SOTA Coding Benchmarks recap — John Yang

Latent Space

AI
Key Takeaways:
  1. The transition from completion to agency means benchmarks are moving from static snapshots to active environments.
  2. Integrate unsolvable test cases into internal evaluations to measure model honesty.
  3. Success in AI coding depends on navigating the messy, interactive reality of production codebases rather than chasing high scores on memorized puzzles.
See full notes
December 31, 2025

[State of Research Funding] Beyond NSF, Slingshots, Open Frontiers — Andy Konwinski, Laude Institute

Latent Space

AI
Key Takeaways:
  1. The center of gravity in AI is moving from closed-door pre-training to open-source compound systems that prioritize context management.
  2. Identify research teams with long histories of collaboration and fund them before they incorporate to capture the highest upside.
  3. Open research is the only way to maintain a democratic and competitive AI ecosystem against both closed labs and international rivals.
See full notes
December 29, 2025

Beyond the Code: The Books That Shaped the Minds of AI Leaders

Turing Post

AI
Key Takeaways:
  1. The transition from technology push to market pull requires builders to stop focusing on the stack and start obsessing over user psychology.
  2. Apply the Mom Test by asking users about their current workflows instead of pitching your solution. This prevents building expensive features that nobody uses.
  3. The next decade of AI will be won by those who understand the human condition as deeply as they understand the transformer architecture.
See full notes
December 29, 2025

Memory in LLMs: Weights and Activations - Jack Morris, Cornell

AI Engineer

AI
Key Takeaways:
  1. The Macro Trend: Moving from "In-Context Learning" to "Weight-Based Memory" to bypass the quadratic costs of attention.
  2. The Tactical Edge: Use synthetic data generation to augment your fine tuning sets and prevent the model from forgetting its base knowledge.
  3. RAG is a stopgap. The long term winners will be those who build "neural file systems" where the model inherently knows the data.
See full notes

Crypto Podcasts

December 17, 2025

Why I’m Shorting $1M of ETH (Again)

Taiki Maeda

Crypto
Key Takeaways:
  1. ETH's current price is likely a function of finite, incentive-driven institutional buying, not organic demand. A significant price correction is probable once this buying pressure subsides, particularly around the January 15th date.
  2. Investors should consider shorting ETH or accumulating cash to prepare for potential market lows. Builders should focus on clear value accrual mechanisms for their own tokens or equity, rather than assuming automatic uplift from underlying infrastructure.
  3. The market is shifting from euphoria to a more rational assessment of value. Understanding the difference between technological utility and asset investment potential is critical for navigating the next 6-12 months.
See full notes
December 16, 2025

How Crypto Networks Are Racing to Be the TradFi Chain - Bits + Bips

Unchained

Crypto
Key Takeaways:
  1. Strategic Implication: The "Empire Strikes Back" is real, with TradFi giants building their own tokenized solutions and specialized chains, intensifying competition for public blockchains.
  2. Builder/Investor Note: Focus on infrastructure and applications that enable seamless movement of tokenized "money" between specialized chains. This interoperability is crucial for unlocking capital efficiency.
  3. The "So What?": Despite current market rotation into "value" assets, the long-term trend of institutional tokenization is accelerating. Regulatory clarity in the US will act as a significant accelerant, but competitive forces are already driving adoption.
See full notes
December 17, 2025

Robinhood’s Onchain Strategy

Unchained

Crypto
Key Takeaways:
  1. Onchain Convergence: Expect more traditional finance players to build on Ethereum L2s, prioritizing security and customizability while abstracting crypto's technical layers.
  2. Tokenization's Reach: The tokenization of private equity and real-world assets will expand, democratizing access and potentially disrupting traditional fundraising and ownership models.
  3. Product-First Crypto: Builders must prioritize user experience and product utility over underlying blockchain mechanics to drive mainstream adoption in the next 6-12 months.
See full notes
December 17, 2025

Lighter vs Hyperliquid: Fees, Fairness, and ADLs - The Chopping Block

Unchained

Crypto
Key Takeaways:
  1. Predictable Risk Management is Paramount: DeFi's long-term success hinges on building transparent, predictable, and fair risk management systems that demonstrably outperform TradFi, especially for institutional players.
  2. Incentive Alignment is Critical: Investors and builders must scrutinize the relationship between DevCo equity and protocol tokens. Misaligned incentives can lead to value destruction for token holders during M&A or other strategic shifts.
  3. The "So What?": The next 6-12 months will see continued innovation in DEX fee models (Lighter's zero-fee tier for retail), RWA derivatives (FX, fixed income), and composability (Lighter's ZKVM sidecar). However, the underlying tension between decentralization ideals and market realities will persist, demanding robust solutions for ADL, governance, and value accrual.
See full notes
December 16, 2025

How Hylo Is Accelerating Solana DeFi In 2026 | Plish

Lightspeed

Crypto
Key Takeaways:
  1. Productive Stablecoins are Key: The transition from unproductive to productive stablecoins like hUSD is a significant catalyst for Solana DeFi growth, attracting capital by offering intrinsic yield.
  2. Builders, Simplify Leverage: Hylo's success with xSOL demonstrates the demand for simplified, liquidation-proof leverage products. Builders should focus on making complex DeFi primitives accessible through intuitive design.
  3. The X-Asset Frontier: Hylo's move into XBTC and other X-assets signals a broader trend: tokenizing leverage for diverse crypto assets will be a major growth driver for DeFi in the next 6-12 months.
See full notes
December 16, 2025

JPMorgan Goes Onchain as the Fed Cuts & Why Bitcoin Can’t Break Out - Bits + Bips

Unchained

Crypto
Key Takeaways:
  1. Institutional Inevitability: Major financial institutions will continue tokenizing traditional assets, creating a clear, low-risk entry point for TradFi into crypto.
  2. Builder Focus: Build infrastructure that bridges TradFi and crypto, or specialize in high-throughput retail solutions. Regulatory compliance and education are paramount.
  3. Market Patience: Expect continued pressure on high-beta crypto assets until a clear market shift occurs, likely requiring high-beta assets to become oversold and the "value" rally to top out.
See full notes