LlamaAI just took a step towards being your personal analyst. Now featuring pptx, docx, and xlsx exports. Create decks, Excel models, and reports from the full set of onchain data.

LLMs are very two-sided for security. On the one hand, long-latent bugs are finally surfacing. On the other hand, before you choose to use a product, you can get your own code review done for incredibly low cost*! In the short term it’s incredibly bad, in the long-term it democratizes code review greatly. “Don’t trust, verify” finally becomes an actual possibility for normal people! * of course LLMs do still make up issues when asked, push any such review to write PoCs to check their analysis!

llama's $PONS revenue now includes the token leg. before only the ETH leg was counted my terminal included i.e. jul 26: llama $387k vs my $166k, ~2.3x. that gap is likely the token leg it was always there just hard to price so no dashboard counted it. when v2 lands those fees get expressed in ETH -> the on-chain numbers should jump too fwiw this is what folks are bullish on from fee revision perspective. hope this helps

Lloyds Banking Group CEO Charlie Nunn explains how he evaluates the return of investment of AI in banking, telling @annaedwardsnews it's still "early days" 

LLMs are all terrible at humor. Maybe in 20 years we’re all just going to be high paid comedians with robot audiences. Imagine selling out MSG and it’s all just Optimus and that weird sock robot with the googly eyes chugging WD-40 and laughing at your jokes. I’m actually totally here for it

LLVM21 is a stinker and has worse codegen in general, and a bug forced disabled loop auto-vectorization. Turned out to be a blessing: we clearly identified areas we were over-reliant on implicit compiler optimizations and were able to explicitly restructure our code to what we really wanted. The result is that we now have clean generic SIMD subroutines in places we previously relied on auto-vectorization. And we did it better and as a result produced faster code (see below). Pure ASCII throughput improved more than 20% (30% on Linux) which is insane and is purely because we wrote better SIMD than an auto-vectorizer can. Lit. We also found it stopped inlining automatically in certain places which destroyed some benchmarks based on real world corpuses. As a result, we now explicitly inline those backed by benchmarks. Wonderful. Some things slowed down more than can be attributed to noise. I'm looking into that now but they're minor benchmarks. The important ones are parity or better.

Llamalend v2 is coming to Ethereum. It brings lending closer to Curve liquidity, with isolated markets, flexible asset pairings and support for Curve LP tokens as collateral. The first markets will roll out gradually through Curve governance.

LLMs are great consensus detectors. @donnelly_brent explains why AI-generated analysis often reflects the institutional average, not a differentiated view. But consensus framing tells you what is already priced and what everyone else is probably looking at.