Daily Edition · Page 8

THE DAILY INDEX

Technology, AI, software, security, crypto and business

HEADLINES & TRENDS

TLDR Founders

PALMER LUCKEY, BACKED BY PALANTIR'S LONSDALE, TO LAUNCH CRYPTO-FOCUSED BANK

Palmer Luckey, supported by Joe Lonsdale, is launching Erebor, a new crypto-focused bank, to fill the gap left by Silicon Valley Bank's collapse. Erebor is seeking a national bank charter and plans to offer services to tech sectors, including AI and crypto, using a digital-only model. The bank, which aims to be the most regulated entity in this sphere, will integrate stablecoins for fast transactions.

AI, SEO, AND HOW TO SIMPLIFY YOUR CONTENT STRATEGY

Apple's Eddy Cue just testified that search volume dropped for the first time in 22 years because people are using ChatGPT instead. Google's stock crashed 7%, wiping out $100 billion. However, here's what everyone's missing: when people ask AI tools buying questions, they're asking the same high-intent queries that drive SEO. "What is the best CRM software?" works the same whether you type it into Google or ChatGPT. Instead of panicking about AI killing search, just answer the questions your sales calls reveal customers actually ask.

GETTING YOUR FIRST AES OFF THE GROUND

Hiring your first AEs before you have a pipeline is like buying a race car without building a track. Founders must prove outbound works themselves first. The biggest mistake is that founders outsource their target account list too early. Your new AEs see logos, you see patterns from dozens of customer calls. Build the list yourself, automate the outreach so AEs start with full calendars, then get out of their way. Their only job is closing deals, not fixing your CRM.

IS THE "CONVICTION" GAME IN VC BS?

Carta ran 20,000 simulations comparing two VC strategies and found something VCs don't want to hear. A "spray and pray" fund with 100 investments beats a "conviction" fund with 20 investments almost every time - 2.6x returns vs 2.1x, and 99.7% chance of not losing money vs 81.9%. Only the top 5% of funds do better with conviction. Yet VCs still concentrate because LPs love the story of genius stock-picking. The math says most VCs would do better admitting they can't predict winners and making more bets.

HOW TO MEASURE AI IMPACT IN ENGINEERING TEAMS

The phrase, "AI improves developer productivity by 100x," is meaningless hype without data. Most engineering leaders report 0.3-1x actual improvement. DX's framework measures what matters - not just if developers accept AI suggestions, but whether that code makes it to production. Track three dimensions that follow the AI adoption lifecycle: First, utilization (are people using it?), then impact (is it actually helping?), and finally, cost (is the ROI positive?). Most importantly, measure AI agents as part of the team overseeing them, not as standalone contributors.

TLDR IndexPage 8