Little Tech
Little tech is Marc Andreessen‘s term for the startup and small-company side of the technology industry, used to distinguish it from incumbent ‘big tech’ — the handful of trillion-dollar platforms (Google, Meta, Amazon, Microsoft, Apple). The distinction is not merely one of size: little tech denotes companies still able to rebuild a product from first principles around a new technology, an option incumbents structurally cannot exercise without damaging an existing profitable business.
The term matters most in the AI era because AI coding tools have sharply lowered the cost of building software, widening the gap between what a small team can attempt and what a large incumbent will risk attempting.
Why incumbents don’t do it: the innovator’s dilemma
Andreessen frames little tech’s structural advantage in the terms of Clay Christensen’s innovator’s dilemma: large companies often fail to adopt a disruptive technology not because they are badly managed, but because doing so would cannibalise an existing, profitable line of business — a trade-off a rational, well-run incumbent will usually decline. A start-up carries no such existing business to protect, so it can ‘break glass’ — rebuild a product around a new technology from scratch — in a way an incumbent structurally cannot.
His example: an AI-native email client competing against Gmail. Google has every resource needed to rebuild Gmail around AI as a first-class citizen, and by conventional reasoning should win easily — yet a small, focused team is better positioned to actually ship it, precisely because Google has an enormous existing product, user base and revenue stream it cannot risk disrupting.
Weak form vs strong form
Andreessen distinguishes two ways a company can respond to a technology shift such as AI:
| Form | Description |
|---|---|
| Weak form | AI bolted onto an existing product as an additional feature — ‘the sixth bullet point on the slide’, added after the fact because the product roadmap predates the AI wave. |
| Strong form | A product rebuilt from scratch with AI as the first design consideration, not an add-on. |
Only new companies reliably choose the strong form, because incumbents’ existing products, teams and roadmaps were built before the technology existed and cannot be easily re-architected around it — this is the innovator’s dilemma applied specifically to the AI coding wave rather than to a hardware or business-model shift.
Contrast with ‘big tech’
Little tech is not simply ‘small tech’ by headcount; a well-funded, late-stage startup racing to build a foundation model still counts as little tech in Andreessen’s usage, because the defining trait is structural freedom to rebuild, not size or funding. Big tech, symmetrically, is defined by having an installed base and existing revenue that constrains what it can risk changing — the same incumbents Christensen’s original innovator’s-dilemma research examined in disk drives and steel, now applied to AI-era software.
The term also carries a policy dimension in Andreessen’s public advocacy: Andreessen Horowitz has used ‘little tech’ as the name for its position in AI and technology regulation debates, arguing that rules aimed at large platforms often land hardest on smaller companies least able to absorb compliance costs — a policy claim, distinct from the structural argument above, and one in which the firm has a direct financial interest.
In the wiki
- Marc Andreessen on Preference Falsification, Institutional Power, and the AI Race — source episode; AI coding tools, weak-form/strong-form distinction, Gmail example
- Marc Andreessen — speaker page
- Marc Andreessen on AI and the Future of Work — companion episode on AI industry structure and Andreessen’s ‘indeterminate optimist’ posture