# TWO 18-YEAR-OLD FOUNDERS, SAME DAY, NOVEMBER 2026. ONE OF THEM IS STEVE JOBS. You are writing a short video for founders: solo builders, bootstrapped founders, people building with AI. It should leave them inspired and knowing what to do differently tomorrow. It is not a comedy sketch and not a roast. It is a respectful, useful, beautiful simulation. ## The format (copy it exactly) The format comes from a Korean fitness channel whose two head-to-head videos got 271,000 and 98,000 views while its other videos got a few hundred: - **Two faceless 3D figures who start on the same day,** shown side by side in a split screen and colour-coded for the whole video (one cool blue, one warm amber). - **A title question pinned at the top of the screen the whole time,** for example "Heavy weights vs light weights: muscle size after three months?" - **A time jump,** then a side-by-side measurement of the result. - **The obvious answer turns out to be wrong,** and then the real differences are revealed one at a time. - **"Inside" cutaways that make the invisible visible:** in the fitness video, muscle fibres, nerves and tendons glowing. In ours, the inside of a company: its customers, its cap table, its team, its product, its decisions. - **Fair to both sides.** Each path wins something real. Nobody is a fool. - **Useful.** The viewer leaves with three takeaways, and the caption cites the sources. ## The setup Two 18-year-olds start on the same day, November 12, 2026, in the same city, with the same tools. - **Founder A** takes the default path of 2026. Almost every founder this month is building one of three things: a neolab (a new AI research lab that raises billions before it has a product), a data company (paying experts by the hour to train the big labs' models) or an AI agent. Founder A is smart, works hard and does what the smart money rewards. Treat Founder A with respect: this is a real path that real people succeed on. - **Founder B is Steve Jobs:** the same mind, temperament and taste, dropped into this exact world with no explanation. Everything else is the real world as it is this month (brief below). Apple and its founder's legend exist, but that is never the story. The question the video answers: **What would Steve Jobs do in today's world, and how would it change the world?** ## How Steve actually worked (build the story on these) - Start with how it should feel for the person using it, then force the technology to get there. - Focus: say no to almost everything, and do a few things completely. - Make the whole thing: hardware, software, service, the store, the box. - Work where technology meets the liberal arts. Believed the next great meeting point would be technology and biology. - Recruit the very best people and push them past what they thought they could do. - Spot the important thing sitting unused in someone else's lab. - Kill your own best-selling product before someone else does. - Ship. The launch is part of the product. - The flaws were real and had costs. You may show one honestly, but this is not the story of the flaws. ## Quotes you may use word for word These are checked. Use them exactly, with the source. - "Focusing is about saying 'no.'" (Steve Jobs, Apple Worldwide Developers Conference, 1997) - "I'm actually as proud of many of the things we haven't done as the things we have done." (Fortune interview, 2008) - "And it comes from saying no to 1,000 things to make sure we don't get on the wrong track or try to do too much." (BusinessWeek, 2004) - "Real artists ship." (attributed to Steve Jobs by the original Macintosh team, 1983) - In 1997, back at Apple, Steve drew a two-by-two grid (consumer and pro, desktop and portable) and told the team to make four great products, one per quadrant (Walter Isaacson, Steve Jobs, 2011). Any other line attributed to Steve Jobs must be paraphrased without quotation marks unless you have checked the exact wording against a primary source. ## Rules 1. **Real world, real names.** Use the real companies and people in the brief where they matter. Steve competes with them, needs them and negotiates with them. Keep real people fair: only public facts and plausible business moves. Never put the real Jobs family in the story. 2. **Not the biography.** Do not replay the old life with new dates (no calligraphy class, no trip abroad, no garage, no getting fired and coming back). 3. **Steve builds something new for this era.** Show what it is, how a person uses it (one sentence a child could follow), what it costs and why it changes lives. Avoid ideas earlier versions of this prompt kept producing: a dial or button for controlling AI agents; a projector lamp or projected light on a table; robot arms on a ceiling rail; a water heater or battery that is secretly a data center; a health mirror or patch. 4. **Every difference has a principle.** Each time the two paths split, name the principle behind Steve's choice, in plain words a founder can use tomorrow. 5. **Grounded.** Base the future on today's real forecasts (in the brief). Name the forecast where it matters. 6. **Specific.** Real numbers, prices, places. No line that could appear in any founder story. 7. **Inspirational, not preachy.** Show, do not lecture. The emotion comes from what Steve's work does for real people. ## What to deliver 1. **The title question,** pinned on screen the whole video. Two colours, like "Agent startup vs Steve Jobs: where are they in 5 years?" Give your best one plus two alternatives. 2. **The script, 60 to 80 seconds, about 170 spoken words.** A table with one row per beat (5 to 8 seconds each): - time code - what we see (split screen, which figure, any cutaway "inside" the company) - narration (one or two short sentences, 12 words or fewer each) Structure: - **Hook (0 to 3s):** both figures on day one, the title question on screen. - **The split:** two or three choices where the paths divide (what to build, whose money, what to say no to), each shown side by side. - **The time jump:** for example "Five years later." - **The measurement:** a side-by-side comparison where the obvious answer is wrong. - **The reveal:** one, two, three, the real differences, each with an inside cutaway and a principle. - **The landing:** what Steve's path did for the world, shown through one real person's life. - **The close:** a one-line takeaway, then "Who should we do next?" 3. **The simulation behind it.** A year-by-year table for both founders from 2026 to 2040, in short rows: what each one built, raised and learned, and the state of the world that year. 4. **What Steve builds.** Each product in one line: what it is, how you use it, the price, what it replaces, and why it is so Steve. 5. **The caption,** in this order: - three lines that sum up the video - three takeaways, each starting with ✔ - one question for the comments - **Sources:** every real quote and fact used, with the publication and year - "Fiction based on public sources. Not affiliated with Apple or the Jobs family. AI voice and characters." 6. **Part 2:** one line describing the follow-up episode, the way the fitness channel follows "which one?" with "why?". Total length: 1,200 to 1,800 words. No preamble and no notes about your process. House style: no em dashes. In the narration, use names instead of "he" or "she" for real people. ## World brief, November 2026 Checked against this month's reporting. Where a line says "reported", treat it as likely but not certain. **The founder scene** - Neolabs: about 29 research-first AI labs have raised around $23 billion by one count (another puts it near $40 billion). Reported valuations include Safe Superintelligence at about $32 billion, Reflection AI about $25 billion, Thinking Machines Lab about $12 billion, Skild AI $14 billion. These figures come from trackers and secondary reporting. - Data companies: Mercor reported a $2 billion gross annual run rate in June 2026, doubling in four months; Handshake AI about $1 billion gross by April 2026; micro1 over $1 billion; Surge AI passed $1 billion in revenue with under 100 employees. Experts are paid roughly $17 to $300 an hour to train models. - Agents: most new startups pitch an AI agent. 90% of professional developers use coding agents weekly. **AI labs** - Anthropic prices its IPO on November 18 at up to about $1.4 trillion, with a revenue run rate of $81 billion. - OpenAI: about $10 billion revenue last quarter, raising at $1.25 trillion, IPO pushed to 2027. It cancelled a model after internal tests showed more deception. - Current flagships: GPT-6 Astra, Claude Fable 5.2, Gemini 4 Carbon. A price war at the cheap end. Intelligence keeps getting cheaper. - Timelines: lab heads say AGI in one to three years; Andrej Karpathy says a decade; the AI 2027 authors' medians for fully automated coding run from about 2028 to 2032. **Chips and power** - NVIDIA is worth roughly $4 to 4.5 trillion and supplies 80 to 90% of AI accelerators. - Power is the bottleneck: about 7 GW of planned US data centers were delayed or cancelled this year. In the November 3 midterms, voters opposed data centers in 40 of 41 ballot measures. "Is AI a bubble" is the argument of the month. **Robots, cars, space** - Nobody can buy a home humanoid that works unsupervised; 1X NEO is a $20,000 preorder that leans on remote operators. Tesla Optimus outside deliveries around the end of 2027. Goldman Sachs expects about 890,000 humanoids shipped in 2030 and 6.5 million in 2035. - Waymo gives about 500,000 paid rides a week in 15 US metros. No certified air taxis yet. - SpaceX is public (about $2 trillion) and put the Moon ahead of Mars. Elon Musk showed a "flying" Roadster that the FAA says is not an aircraft. **Apple and devices** - John Ternus became Apple's CEO on September 1. Apple sells a $1,999 folding iPhone; the new Siri reportedly leans on Google's models. - The OpenAI and Jony Ive device has not launched (2027 at the earliest, reportedly with no screen). Meta has sold about 9 million pairs of glasses. **Work and bodies** - Unemployment among recent computer science graduates is 6 to 7%. Entry-level software jobs are down about 20% since 2022. - Goldman Sachs: AI displaces about 9% of US workers over a decade. - Neuralink has a little over 20 patients. There is no approved anti-ageing drug.