Mert says Zcash gave NEAR momentum, but NEAR may have helped unlock Zcash first “You have to look at what outperforms the general market, which Zcash and NEAR have done” “If you just say it's encrypted Bitcoin, that's pretty much all you need” “It's questionable that Zcash would have went up as much if it weren't for NEAR in the first place” @mert thesis is a two-way flywheel. Zcash brings a simple monetary + privacy narrative. NEAR Intents gives $ZEC more utility by connecting it to wallets, swaps and liquidity across other chains. More access creates more ways to use $ZEC beyond simply holding it on an exchange, while Zcash activity feeds back into the $NEAR ecosystem. @Zcash brings the asset. @NEARProtocol expands where it can move and what it can do.

Introducing The Latent, The Block's sister publication covering AI. Launching today: AI news, a data portal, The Latent Score, and a daily newsletter. We're bringing our reporting and data expertise to AI. See what we're building and subscribe below.

Pi beats claude code and codex You can cut your agent API costs by 50% right now with literally 0% drop in benchmark performance. Anthropic wants you to use Claude Code. OpenAI wants you to use Codex CLI. But a new UC Berkeley study just exposed the "In House Myth": Frontier models actually perform better on minimal, open source frameworks 75% of the time. Stop paying the Big Tech "Harness Tax." Here is the data # UC Berkeley and Arena just dropped the "Harness Tax" paper. They benchmarked 21 model agent pairs on SWE-bench Lite and Terminal-Bench 2.0. They tested SOTA models (Claude Fable 5, GPT-5.6 Sol, Claude Opus 4.8, Sonnet 4.6, Kimi K3) across 3 agent frameworks: Claude Code, Codex CLI, and a barebones open source harness, Pi. The results prove we are all bleeding API credits for literally zero reason. # Look at the chart: - GPT-5.6 Sol scored 83.3% using the OS Pi harness. On OpenAI's own Codex CLI? It dropped to 78.9%. and cost nearly 2x as much ($0.42 vs $0.76). - Claude Fable 5 hits 97.8% on Claude Code for $1.33/rollout. On Pi? It hits 96.7% for just $0.67. You are paying a 100% markup for a 1% gain. # Why is this happening? First Call Bloat. Proprietary coding agents load up massive system instructions, heavy guardrails, and verbose JSON tool schemas. Claude Code sends 10x the initial context of Pi. You are paying for their dead weight context on every single turn. (check out the replies for more data on this) Pi gives the model exactly 4 tools: read, write, edit, bash. That's it. The takeaway: As frontier LLMs get smarter, heavy agent scaffolding actually holds them back. If your agent API bill is too high, do NOT downgrade your model. Just ditch the bloated harness. I've dropped more data and link to the study in the replies. What agent framework are you guys running in your dev loop right now? Let me know

I need this real bad.

After co-inventing ChatGPT, I kept asking myself: why have superhuman chat models not led to AGI? I’ve spent the last 2 years in stealth building a new way to train models (RLCD), and a new type of frontier AI model that we are releasing today: Jev • 20-200x faster • 40-400x cheaper (w/ output tokens free) • Frontier composable intelligence optimized for decisions AFAICT the shortest path to AI-based economic revolution

if you’re raising for a crypto project, these are active investors worth talking to: - @ellazhang516 (YZi Labs, 30 deals in the last 6 months, more than any other fund, also runs the EASY Residency for founders) - @HoolieG, @jonathankingvc (Coinbase Ventures, 29 deals, second most active) - @eddylazzarin (a16z crypto, $2.2B Fund V in May, 17 deals) - @rainandcoffee (Maven 11, tier 2, 9 deals, Europe) - @HadickM, @hosseeb (Dragonfly, $650M Fund IV in February, oversubscribed) - @matthuang (Paradigm, $1.2B fourth fund in July) - @katie_haun (Haun Ventures, $1B across two funds in May) - @jessewldn (Variant, $222M for Variant Fund IV) - @KyleSamani (Multicoin) - @CarlKVogel, @mdudas (6MV) - @collectorofgems, @RealMissAI (MH Ventures, 149 deals) - @dara_venture (deploying out of a first close while still raising) - @reganbozman (Lattice, $200–500K first checks, raising Fund III) - @QwQiao (Alliance, the accelerator, $250K standard check) - @rleshner (Robot Ventures, $100–500K, also runs Superstate) - @ZeMariaMacedo (Delphi co-founder) - @dcfgod, @balajis, @cory, @LucaNetz, @santiagoroel (active angels) who am i missing?

Anyone know someone @expo? Need access to their iOS simulator real bad.

does your ai agent have brain damage? you might be flooding the context... even if your model has a 1M context window it gets the best work done 100k - 200k beyond 300k context all bets are off @kain