Hebbia

Lighthouse or Landgrab: Choosing an AI Sales Strategy

The article contrasts two go-to-market playbooks for AI enterprise companies: the 'Lighthouse' strategy (winning marquee customers to provide social proof for category-creating products) and the 'Landgrab' strategy (moving fast to win many customers on clear ROI when buyers already understand the problem). It explains how buyer exposure and whether social proof 'travels' determine which approach fits a given market, illustrates each strategy with examples (Harvey, Hebbia, Stuut, Decagon, Affirm), and warns of common pitfalls — pilot purgatory, vanity logos, over-customization, and scaling before product readiness. The piece recommends sequencing from lighthouse to landgrab where appropriate and using buyer risk/reward calculus to choose the right sales motion.

a16z launches Talent Engineer Fellowship

Andreessen Horowitz (a16z) has launched the a16z Talent Engineer Fellowship, an 8-week cohort program aimed at engineers who build recruiting systems and workflows (termed 'Talent Engineers'). The fellowship targets builders who create sourcing agents, candidate-intelligence workflows, relationship graphs, AI-assisted technical evaluations and other talent-infrastructure tools. Applications are open on a rolling basis with an application deadline of July 17; the program begins in late August. The cohort will be a tight network of roughly 40 AI-native talent leaders, include in-person dinner events in San Francisco (with expansion to other cities by demand), demos, working sessions, and access to the a16z ecosystem. Founding fellows and industry participants include named talent leaders from companies such as Cursor, OpenAI, xAI, Ramp, Zapier, Anthropic, Airbnb and Luma AI.

OpenAI Unveils Fast, Affordable GPT-5.4 Mini and Nano

OpenAI announced GPT‑5.4 mini and GPT‑5.4 nano, smaller, faster variants of GPT‑5.4 optimized for high-volume, low-latency workloads such as coding assistants, subagents, computer-usage tasks, and real-time multimodal applications. GPT‑5.4 mini delivers substantial gains over GPT‑5 mini in coding, reasoning, multimodal understanding, and tool use while running more than 2x faster; GPT‑5.4 nano is positioned as the cheapest, smallest option for classification, data extraction, ranking, and simple coding subagents. GPT‑5.4 mini is available in the API, Codex, and ChatGPT (400k context window); GPT‑5.4 nano is available in the API only. OpenAI published benchmark comparisons and pricing: GPT‑5.4 mini ($0.75 per 1M input tokens; $4.50 per 1M output tokens) and GPT‑5.4 nano ($0.20 per 1M input; $1.25 per 1M output).