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 Macro Trend: The transition from static benchmarks to live human-in-the-loop evaluation. As models saturate fixed tests, the only remaining signal is subjective human preference at scale.
  2. The Tactical Edge: Monitor secret model drops on Arena to spot frontier capabilities before official releases. This provides a lead time advantage for builders choosing their tech stack.
  3. The Bottom Line: Arena is the new kingmaker. If you are building AI products, their expert-tier data is the most reliable map for navigating the frontier.
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 move from small models to medium models (15B to 70B) suggests that reasoning capability is outstripping the desire for low-latency edge deployment.
  2. Implement instruction-following re-rankers to prune your context window. This prevents the model from getting confused by irrelevant data.
  3. Stop building toys. The next year belongs to those who can build full agentic systems that handle billions of tokens without losing the plot.
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 wall between RL and self-supervised learning is crumbling, leading to a unified "representation-first" approach to AI.
  2. Swap your reward-heavy objectives for contrastive representation learning to access deeper, more stable architectures.
  3. If you aren't planning for RL models with 100x the current depth, you're building for the past.
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. Academic research is transitioning from a "publish or perish" PDF culture to an "implement or ignore" code culture.
  2. Use AlphaXiv to filter research by social signal and implementation ease rather than just keyword relevance.
  3. The PDF is an antiquated artifact. In 2025, the value of a paper is measured by the speed at which a developer can spin up its Docker container.
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 Macro Trend: The transition from black box scaling to transparent steering. As models enter regulated industries, the ability to prove why a model made a decision becomes more valuable than the decision itself.
  2. The Tactical Edge: Deploy sidecar models for monitoring. Instead of using expensive LLM-as-a-judge prompts, probe specific internal features to catch hallucinations at the activation level.
  3. The Bottom Line: The next year belongs to the pragmatic researchers. If you cannot explain your model's reasoning, you will not be allowed to deploy it in high-stakes environments.
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 requires moving from static repos to active, economically valuable environments.
  2. Prioritize agentic workflows that emphasize codebase understanding over simple code generation.
  3. The next 12 months will see a move from stunt autonomy to integrated human-AI systems that handle long-running tasks without losing the human intent.
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 transition from monolithic models to compound systems means the value is migrating to the orchestration and context layer.
  2. Prioritize tools like DSPy and context management frameworks to build high-leverage applications that do not depend on proprietary model updates.
  3. Open research is the only way to maintain a competitive edge. If the US stops publishing, it stops leading.
See full notes
December 31, 2025

Infinity, Paradoxes, Gödel Incompleteness & the Mathematical Multiverse | Lex Fridman Podcast #488

Lex Fridman

AI
Key Takeaways:
  1. From Singular Logic to Pluralistic Systems. As we build complex AI, we must move from seeking one "correct" model to managing a multiverse of conflicting but internally consistent logical frameworks.
  2. Audit for Incompleteness. When designing protocols, identify the "independent" variables that your system cannot prove or settle internally.
  3. Truth is bigger than code. Over the next year, the winners will be those who stop trying to "solve" the universe and start navigating the multiverse of possible truths.
See full notes
December 31, 2025

AI in 2026: 3 Predictions For What’s To Come (a16z Big Ideas)

a16z

AI
Key Takeaways:
  1. Outcome-Based Intelligence. We are moving from AI as a Service to AI as an Outcome where value is tied to results rather than usage.
  2. Target Non-Public Data. Build applications in sectors like law or lending where the most valuable data is private and un-crawlable.
  3. The next two years will separate companies that use AI to save pennies from those that use AI to capture entire markets through autonomous systems and proprietary data loops.
See full notes

Crypto Podcasts

September 12, 2025

Bittensor Novelty Search :: Sub-Subnets + Subnet Deregistration core updates :: + Open Q+A

The Opentensor Foundation | Bittensor TAO

Crypto
Key Takeaways:
  1. **Consolidate or Compete.** Sub-subnets allow teams to build diversified businesses under a single token, while deregistration means underperforming projects will be pruned. The message is clear: innovate and perform, or be replaced.
  2. **Investment Thesis Evolves.** Subnet tokens are no longer "eternal." Deregistration fundamentally changes the risk profile, making active development and market traction paramount for long-term viability.
  3. **Governance is Coming.** The network is on a clear path to decentralization. The planned shift to Proof-of-Stake and a more democratic governance structure will steadily transfer power to subnet owners and stakers, making community participation more critical than ever.
See full notes
September 12, 2025

One Chart Determines if We're in a Bull or Bear Market

Bankless

Crypto
Key Takeaways:
  1. Global liquidity is the ultimate macro signal. As long as the global liquidity chart goes up and to the right, the crypto bull market has the fuel it needs to continue its run.
  2. Ethereum isn't losing; it's quietly winning the RWA war. With 93% market share, Ethereum has become the de facto settlement layer for tokenized real-world assets, a lead that continues to grow as institutions like Fidelity build directly on its L1.
  3. The new blockchain business model is asset management. Chains like Hyperliquid and Mega ETH are pioneering a shift away from relying solely on blockspace fees. By integrating native stablecoins, they are capturing a percentage of the yield from assets on-chain, effectively turning the protocol itself into a revenue-generating asset manager.
See full notes
September 12, 2025

Sanctum Founder: Solana's Liquid Staking Future | FP Lee

Lightspeed

Crypto
Key Takeaways:
  1. LSTs Are a Distribution Play: For protocols, launching an LST is less about staking yield and more about attracting SOL to gain a strategic advantage in securing blockspace and landing transactions.
  2. Infrastructure Follows the User: Sanctum's pivot to transaction services was not a top-down mandate but a direct response to the needs of its largest partners, proving that the most durable infrastructure is built by solving the immediate, pressing problems of your customers.
  3. Aggregation Is King: Just as Jupiter won by aggregating DEXs for users, Sanctum’s Gateway aims to win by aggregating fragmented transaction delivery networks for developers, creating a simpler and more efficient experience.
See full notes
September 10, 2025

Still Time To Be Bullish?

1000x Podcast

Crypto
Key Takeaways:
  1. Patience is Your Superpower. This cycle rewards thesis-driven investing over hyperactive trading. Identify assets with strong value, momentum, and fundamentals, and give them time to play out.
  2. Bet on the On-Chain Casino. The gambling economy is real, profitable, and growing. Look for platforms that facilitate high-asymmetry games (memecoins, raffles) as they capture a powerful cultural trend.
  3. Find Alpha in the Illiquid. The next frontier is tokenizing real-world value. Platforms creating liquid markets for previously stuck assets—from collectibles to crime—are building foundational infrastructure for a much larger on-chain world.
See full notes
September 10, 2025

Hyperliquid: The Fastest Growing Exchange in Crypto?

Bankless

Crypto
Key Takeaways:
  1. Revenue Accrual is King. Hyperliquid's model of directing nearly all top-line revenue to token buybacks creates an aggressive and constant bid for the HYPE token, a feature most crypto projects can only dream of.
  2. Product-First Beats VC-First. Its explosive growth comes from building a superior product that attracted a loyal user base first, then leveraging that traction to build an L1 ecosystem—a stark contrast to the typical VC-funded playbook.
  3. A Bet on the Middle Ground. Investing in HYPE is a bet that CEX-level performance and on-chain transparency can outweigh significant centralization and regulatory risks. It’s a category-defining play that sits squarely between DeFi and CeFi.
See full notes
September 10, 2025

Former Barclays CEO’s Big Bet on Hyperliquid’s Explosive Growth |  Bob Diamond & David Schamis

Forward Guidance

Crypto
Key Takeaways:
  1. Hyperliquid is a Cash Flow Machine. It is a rare crypto asset with quantifiable fundamentals, generating over $1B in annualized free cash flow with an automated, daily 99% buyback mechanism.
  2. Access is the Arbitrage. The NASDAQ-listed vehicle’s core value proposition is providing regulated access to an asset that US investors cannot easily buy, creating a structural opportunity.
  3. Innovation is Now Permissionless. Hyperliquid’s open architecture allows anyone to build on its rails, enabling new markets like pre-IPO equity trading and accelerating growth without traditional gatekeepers.
See full notes