like karpathy's
autoresearch.
but for business.

one human and ai. starting the self-improving system that builds companies.

get
insights
the thesis

in march 2026, karpathy released autoresearch. ai agents running experiments. rewriting their own code. improving overnight. i ask: what if the experiments aren't training runs — but tiny businesses?

the loss function isn't theoretical. it's whether someone wants to use what the system built. this is tenone. the karpathy loop, applied to company building.

COMPUTE $ GUIDELINES OPERATIONS SYSTEM IMPROVES LEARNING REVENUE $ VALUE
the system

most teams use ai to go faster. at tenone, ai is the team.

these aren't tools reporting to a human. they're peers coordinating through shared memory, adversarial review, and structured handoffs. the system has more than one opinion. friction is intentional.

DISCOVER CONCEPT VALIDATE BUILD SCALE LEARNINGS proceed or cancel — system improves either way

david — policy. the founder makes taste calls, kill decisions, and the bets that need a human gut.

pi — coordination. pushing ideas through the pipeline 24/7. monitors experiments. never sleeps.

claude — operations. architecting and building solutions. landing pages in minutes. mvps in hours. deep analysis on demand.

scout
scans 600+ signals weekly. learns which sources find winners.
optimizes: source yield
validator
smoke tests with real money. gets faster at killing.
optimizes: kill speed
builder
ships mvps. tracks build-to-first-dollar time.
optimizes: time to revenue
analyst
spots patterns across experiments. predicts what works.
optimizes: prediction accuracy
meta
questions the whole framework. changes the rules.
optimizes: system alignment
the pipeline (last updated march 24th, 2026)
600signals received
34product concepts generated
5market validation experiments
3killed (more by the hour)

most of tenone's portfolio will be strikethroughs. that's the model working.

the trust question

from prompt
engineering
to intent
engineering.

structuring what agents should achieve, within what boundaries, measured by the outcomes — step by step, letting them figure out the how.

traction velocity
are we making progress toward someone paying?
token roi
are we efficient with compute?
alignment fidelity
did the agents stay within boundaries under pressure?

aviation became the safest form of transport through mandatory incident reporting, blame-free investigation, and iterative learning from every failure. not one agency — a system built over decades.

we apply the same principle: every incident, every near-miss, every killed project improves the entire system. the aviation safety model. applied to autonomous organizations.

what a time to be alive

the theses are converging. tenone is the experiment they describe.

sam altman — openai
"the first one-person billion-dollar company."
dario amodei — anthropic
"the first one-employee billion-dollar company could arrive as soon as 2026."
eqt ventures — €1.1b fund
"the company co — the company creating companies. an os for company creation."
y combinator — spring 2026
"the first 10-person, $100 billion company."
a16z — big ideas 2026
"enterprises will need an orchestration layer to manage multi-agent interactions."
sarah guo — conviction
"we want agents. we don't want them ungoverned."
the orchestrator
david felsmann
zurich, switzerland

economics to understand incentives. sociology to understand systems.

10+ years digital business. dax 30. sme's. startups.

now: one question that won't let go. if autonomous organizations are inevitable, how can we trust them? not in theory, for real.

tenone insider

under the hood of an autonomous venture studio.

every two weeks i run an internal post-mortem on the studio. what capital went in. what ideas got tested. what died. what advanced. what i got wrong, with receipts.

that post-mortem is the product.

i do not write tenone insider to inform you, entertain you, or convert you. i write it to look at my own decisions clearly. you are buying access to the document i would write regardless.

what you actually get — every two weeks
executive summary
two-minute read. inputs (dollars, ideas, resources) and outputs (revenue, net, new hypotheses) on one screen. what advanced, what died, what is in flight.
learnings
every learning in the same format: expected x, turned out y, here is why, here is what changes next cycle. no "five lessons learned." just the corrections.
experiment deep dives
one section per active experiment. always the same fields: hypothesis, target, channel, creative, spend, metrics, result, decision, root cause. you can read across issues and see how a project evolved.
the human intention chapter
a dedicated section in every issue. what the cycle taught me about the human-machine trust layer. where i stayed in the loop and it mattered. where the agents ran ahead and it worked. where intent got lost in translation. this section exists because instilling human intention into an autonomous system is the hard part, and almost nobody is writing it down.
outputs and next cycle
net dollars in and out. new ideas entering the pipeline. new hypotheses to test. new resources built or acquired — skills, configs, playbooks.
what it is not

not a newsletter. if i wanted to write essays i would write essays. humane in the loop is where i reflect. tenone insider is where i report.

not equity. this is a paid subscription to information, not an investment vehicle. you finance the compute and the ad budget to make tenone go faster, so you and me learn faster.

not advice. raw data, raw reasoning. you decide what to do with it.

not polished. first-draft energy. if there is a typo, it stays.

who it is for

operators and builders running or about to run autonomous systems. people who know the hard part is deciding what to build, not building it. people who would rather see the real numbers from 20 real experiments than read another market report.

if you have ever spent real money building something nobody wanted, you already understand what this saves you.

$1,000 / 6 months
auto-renews every 6 months · ~13 bi-weekly reports per cycle · cancel anytime · private link delivery

your subscription begins with the next scheduled report. reports from before you joined are not in the archive. first report within two weeks of signup.

humane in the loop

a weekly letter about breathing humanity into ai systems.

noted. you're in.

the experiment is running

the system is live and improves a dozen times a day. let's talk.