Guide

The Machine That Runs Dessa: MacBook Pro M4 Max Guide

2026-11-04 · 7 min read · By DessaAI Daily

Disclosure: some links in this article may earn us a commission at no cost to you if a programme is active; where no programme is confirmed, links are plain product references with no affiliate code.

One of Block's most quoted Dessa claims is not about conversation at all. It is about hardware: inference runs faster than real time, locally, on a MacBook Pro with an M4 Max chip or better. That single line turns an abstract research preview into a concrete shopping question — what exactly is the machine that runs Dessa, what does it cost, and should anyone buy one for Dessa today? The honest answer to that last part is no, not yet — and this guide explains why, and what to check if you want one for other reasons.

What Block actually claimed

Block says Dessa was trained on a single NVIDIA GB200 NVL72 system and that the finished model runs locally, faster than real time, on a MacBook Pro with M4 Max or better. Block has not announced a public download, a product form, phone support, pricing or a release date, so “runs on an M4 Max MacBook Pro” is a demonstrated capability in Block's hands, not a shipping promise in yours. There is no Dessa file to install on any laptop today. Our local AI explainer covers why local inference matters for privacy and cost if it does reach users in that form.

What an M4 Max MacBook Pro is

Apple's M4 Max MacBook Pro comes in 14-inch and 16-inch sizes. Retail listings checked today describe the base M4 Max configuration with a 14-core CPU, a 32-core GPU, 36GB of unified memory and 1TB of storage, with higher configurations offering more cores, memory and storage. That unified memory matters for local AI: the model, the audio processing and the operating system all draw from the same fast memory pool, rather than shuttling data between separate system and graphics memory.

For context, listings checked today put a 14-inch M4 Max model with 36GB and 1TB around $3,099 and a 16-inch equivalent around $3,399 at a major US retailer, with higher-memory 16-inch configurations near $3,899. Prices move, retailers differ, and you should always check Apple's own store and reputable retailers on the day you buy rather than treating any guide's snapshot as a quote.

14-inch or 16-inch for local voice AI?

For the Dessa claim itself, screen size is irrelevant — Block's threshold is the chip, M4 Max or better, not the display. The choice is practical. The 14-inch is lighter and easier to carry to a studio, office or café, which suits creators recording voice work on the move. The 16-inch gives a larger display, a larger chassis and, in listings checked today, longer rated battery life claims, which suits a desk-based setup running long local inference or editing sessions. Either satisfies Block's stated floor, because either can be configured with M4 Max.

Memory deserves more attention than screen size. If you are buying a machine to experiment with local models generally — Moshi-based systems and other open voice models you can actually download today, as in our Dessa alternatives guide — more unified memory gives you more headroom for larger models and for running a model alongside recording and editing software. Storage is easier to supplement externally; memory on Apple silicon cannot be upgraded after purchase, so buy the memory you need at order time.

Should you buy one for Dessa?

Not for Dessa alone. Dessa is a research preview. There is no download, no API and no date, and early access goes first to a small group of AI safety researchers through the waitlist at block.xyz/dessa, as our availability guide explains. Buying a $3,000-plus laptop today in the hope of running a model that has not been released would be speculation, not planning, and Block has not promised that a future public release will use this same local form.

Buy an M4 Max MacBook Pro if you already need that class of machine — for video editing, music production, software development, or experimenting with the open local models that exist right now — and treat “it clears Block's stated Dessa floor” as a useful bonus and a sign of where local AI hardware expectations are heading. That framing also explains why Block chose this chip as its proof point: if a speech model this conversational can run on a laptop professionals already own, the economics of voice AI stop depending entirely on per-minute cloud bills.

The bottom line

The machine that runs Dessa, in Block's demonstration, is a MacBook Pro with an M4 Max chip or better — a machine that starts, in listings checked today, at roughly three thousand dollars in the US for the base M4 Max configuration. It is a real, buyable computer. Dessa, for now, is not a real, buyable product. Keep those two facts separate, join the free waitlist if you want access, and let the first independent hands-on reports — not a hardware purchase — tell you when Dessa is ready for your desk.

Sources: Block's official Dessa page (block.xyz/dessa) for the M4 Max inference claim; US retail listings for MacBook Pro M4 Max configurations checked November 4, 2026 — prices are snapshots, not quotes, and should be re-checked at purchase.

Frequently asked questions

← More Dessa AI articles Join the official waitlist ↗