Tech Videos

Here is today’s technical breakdown of the six videos in your daily digest. As engineers, our time is short and our skepticism is high, so I have cut through the marketing noise to focus purely on architectural realism, empirical benchmarks, and tactical execution.

Watch First

How to close $100K+ enterprise deals, step by step | Jen Abel (from Lenny’s Podcast) is the single most valuable video to watch today. It completely demystifies the B2B enterprise sales pipeline by replacing typical, rigid CRM-driven scripts with a highly tactical, 15-step project-management framework that focuses on active listening, uncovering executive mandates, and co-authoring bespoke solutions with technical champions.

Highlights by Theme

Developer Tools & Platforms

In local.ai: hardware benchmarking for local AI models on GitHub’s channel, a new benchmark tool is demonstrated that cuts through raw, vanity “tokens per second” metrics to evaluate local LLMs based on task completion quality, execution speed, and actual hardware and energy costs. Meanwhile, on The Pragmatic Engineer channel, “Treat computing as an onion" urges engineers to peel back high-level abstraction layers to understand JavaScript engines, network requests, compositing, memory, and GPUs. This “mechanical sympathy” allows developers to optimize systems for resource-constrained edge environments where high-level frameworks fail.

AI & Machine Learning

On Dwarkesh Patel’s channel, Redwood Research’s Ryan Greenblatt takes a blunt, conceptual swipe at LLM theory in Machine Learning Is Shallow Compared to Math - Ryan Greenblatt, arguing that ML is a conceptually shallow field. He posits that the field’s most celebrated formulas, like scaling laws, are ultimately simple heuristics that can be explained in minutes, contrasting them with math’s genuinely deep, hard-to-understand abstract frameworks. This suggests that despite the hype around complex “AI science,” ML remains a fundamentally empirical, brute-force engineering discipline rather than a conceptually deep theoretical science.

Hardware & Infrastructure

The All-In Podcast channel tackles the geopolitical bottlenecks of physical compute in David Sacks Calls Out Dario’s Contradictions on AI & Data Centers, where David Sacks highlights the clash between building infrastructure and regulatory gatekeeping. Sacks criticizes Anthropic CEO Dario Amodei for lobbying to block chip exports to protect US-based “data centers full of geniuses” while simultaneously supporting policies that restrict chip sales to allies and overregulate domestic data center construction. Sacks argues that this hypocritical regulatory doomerism effectively blocks the massive infrastructure development required to actually win the AI scaling race.

Everything Else

In Why They Rejected 31,831 PM Applicants on Lenny’s Podcast, Whatnot reveals their brutal hiring filter, highlighting that “politics-focused” and “alignment-building” PMs who merely babysat bloated processes at big tech are an active anti-pattern. Instead, successful product management requires decision-makers who display tight macro and micro execution and act as actual agents of change. Additionally, in her masterclass on Lenny’s Podcast, GM of Enterprise Sales Jen Abel outlines her 15-step sales cycle, emphasizing that 90% of reps fail because they follow rigid CRM scripts instead of actively listening to uncover C-suite mandates and co-authoring pilots with champions.

📊 I can synthesize Jen Abel’s 15-step sales playbook into a clean, structured cheatsheet for your team to use as a quick reference when evaluating sales pipelines or hiring technical managers.