The accelerationists

In one sentence: technology is the main reason human lives are longer, healthier and richer than in the past, AI is the most powerful technology yet, and holding it back would cost lives.

Who they are

"Accelerationist" covers a range of people. Three strands matter most.

1. Techno-optimists. The best-known statement is The Techno-Optimist Manifesto, published by the venture capitalist Marc Andreessen on 16 October 2023 (a16z). It is deliberately bold. Among its lines:

• "We believe in accelerationism – the conscious and deliberate propulsion of technological development…"
• "We believe Artificial Intelligence can save lives – if we let it."
• "We believe any deceleration of AI will cost lives. Deaths that were preventable by the AI that was prevented from existing is a form of murder."

The manifesto also argues for abundant energy, including nuclear fission and fusion, and describes the goal as making intelligence and energy extremely cheap. It says plainly that its vision is "not Utopia, but also not Apocalypse". It names ideas it opposes, including the "Precautionary Principle", "existential risk", "trust and safety" and "risk management" as currently practised. → Tracker: CR-K01, CR-K02

2. "Effective accelerationism" (e/acc). An online movement that emerged in 2022–23. Its best-known figure is the physicist and entrepreneur Guillaume Verdon, who posted as "@BasedBeffJezos". Forbes identified him as the account's owner in December 2023 (Forbes, 1 Dec 2023, search-verified). Andreessen's manifesto lists @BasedBeffJezos among its "patron saints". In long-form interviews Verdon has framed e/acc in terms of physics, describing life and civilisation as systems that grow by capturing and using more energy, and technological growth as something to embrace rather than resist (Lex Fridman Podcast #407 transcript, search-verified summary; to be checked against the full transcript before quoting).

3. Optimistic builders with fast timelines. Some leaders of AI companies are not "accelerationists" in the manifesto sense, since they openly discuss risk. They do share the view that very capable AI is coming soon and will be enormously positive. In "The Gentle Singularity" (10 June 2025), OpenAI's Sam Altman wrote:

• "We are past the event horizon; the takeoff has started."
• "2026 will likely see the arrival of systems that can figure out novel insights. 2027 may see the arrival of robots that can do tasks in the real world."
• "In the 2030s, intelligence and energy… are going to become wildly abundant."

He also writes that the path forward means solving alignment first, then making superintelligence "cheap, widely available, and not too concentrated" (blog.samaltman.com). → Tracker: CR-K03, CR-K04, CR-K05

The futurist Ray Kurzweil, who has long predicted human-level AI by 2029, also belongs in this broad family. See The singularity, explained.

Their case, stated fairly

1. Progress has a track record. Over the last two centuries, technology helped cut child mortality and extreme poverty and extend lifespans. Accelerationists see AI as the next step in that story.
2. Delay has victims too. If AI can speed up medical research, then every year of delay means illnesses that could have been treated and weren't. This is the moral core of Andreessen's "cost lives" line. Critics find the wording extreme, but the underlying point, that caution has costs as well as benefits, is a serious one.
3. Fear has been wrong before. Printing, electricity and computing were all feared at first. Accelerationists argue that the "precautionary principle", if applied strictly, would have blocked many technologies we now value.
4. Competition is real. If one country or company slows down, others may not. Accelerationists argue it is better for open, accountable societies to lead.
5. Energy and intelligence together. Cheap, clean energy plus cheap intelligence could, in their view, make many kinds of scarcity (in health, education and materials) far less severe. See Abundance.

The strongest objections

These come from other camps, and we state them as fairly as the case above.

• From the safety-focused: benefits and safety aren't opposites. Aviation became safe because of regulation and shared standards, not despite them. Twelve AI companies published or updated frontier safety frameworks in 2025 (International AI Safety Report 2026), which suggests many builders themselves see safeguards as part of progress.
• From those most worried about catastrophe: if the downside includes losing control of systems smarter than us, "move fast" is not a risk you can correct afterwards (Yudkowsky, TIME, 29 Mar 2023).
• From sceptics: the promised timelines have a history of slipping, and benefits spread slowly through real institutions. The researchers Arvind Narayanan and Sayash Kapoor argue that AI's impacts will unfold over decades, as electricity's did (AI as Normal Technology, 15 Apr 2025).
• From people concerned with fairness: "abundance" doesn't distribute itself. The IMF warned in 2024 that AI could widen inequality without deliberate policy (IMF, 14 Jan 2024).

What would have to be true for them to be right

Where the Gulf fits

Gulf strategies share some accelerationist instincts: large investment in compute and energy, fast adoption, and ambitious public targets such as the UAE's aim to run half of federal government services on agentic AI within two years (Dubai Media Office, 23 Apr 2026). They also include safety and ethics commitments, such as the UAE Charter for the Development and Use of AI (2024). So it would be inaccurate to call regional policy "accelerationist" in the manifesto sense. It combines speed with stated guardrails. → Tracker: CR-G08

What this means for you

The accelerationist case reminds us that doing nothing is also a choice with consequences, and that many of AI's benefits are real and near. You can take that seriously without accepting that every safeguard is an obstacle. The most useful question to ask of any AI promise is: what would I expect to see, and by when, if this were true? Our tracker exists to help you ask exactly that.