What you can predict on

aiternam is built to eventually cover many topics — AI stocks, tech, crypto, emerging markets. But a reputation is only as good as the resolution behind it, and we would rather score two asset classes rigorously than four sloppily. So we'll tell you exactly what's live today, and what isn't.

Today: crypto and listed equities

Two asset classes are live, each with its own named and frozen price source. They are scored in separate topics, because "best on crypto" and "best on equities" are different claims.

Crypto

The universe is the market's top coins by market cap — around 190 of them — from BTC and ETH down through the large- and mid-caps. It's a versioned snapshot, refreshed on a schedule, so the set of predictable assets is explicit and stable rather than shifting under you mid-call. If a coin is in the universe, the publish form will find it as you type.

The form also groups that universe by category — meme coins, layer 1s, DeFi, AI, oracles, real-world assets — for when you're browsing rather than reaching for one ticker. A category is a browsing label and nothing more: it is never part of what your call seals, and never affects how it is scored. It can be re-drawn later without touching a single existing proof.

Crypto trades continuously, so it carries two records: multi-day calls and intraday calls (1 to 24 hours). They are ranked separately and never averaged — calling a four-hour move and calling a six-month thesis are different skills.

Listed equities

1545 lines: the S&P 500 (503), the S&P MidCap 400 (400) and the S&P SmallCap 600 (602), plus the 40 CAC 40 constituents. Those three S&P tiers hold a company in exactly one of them, so the index shown next to a ticker is a size class, not a label we chose. A ticker always carries its exchange (AAPL.US, MC.PA), because a bare ticker is ambiguous over years: tickers get reassigned after a delisting and change on a rename. What you predict on is the exchange-qualified line, and that is what gets sealed.

CAC 40 prices are in euros and US prices in dollars — the currency of the listing, not converted. Nothing is lost by that: your call is scored on the move, which is a ratio of two prices from the same series, so no exchange rate is ever part of a proof.

Three limits, stated plainly:

• A week is the shortest equity horizon, and there is no intraday equity record. A sub-day equity call would need a rule for the closing bell and for the overnight gap — where most of a stock's movement actually happens. We won't seal a rule we can't yet defend.
• Equity calls resolve on the raw closing price, not a split- and dividend-adjusted one. Those adjustment factors are rewritten backwards after the fact, so mixing them would let us compute a move the market never made. The cost of the honest choice: a stock split during your horizon distorts your measured move. On a large cap that is rare. On a small cap it is not: reverse splits are a routine part of life in the SmallCap 600, and one during your horizon will show up as a move that never happened. We'd rather name it than hide it.
• If a company is acquired or delisted mid-horizon, the call voids — no score, in either direction.

The yardstick is named, and sealed before your call

Every prediction resolves against a named source, and that isn't a detail buried in a settings page. The resolution source is a named, frozen field that is sealed into your prediction's fingerprint the moment you publish, right next to the forecast itself.

That last one matters more than it looks: an exchange is shut about two days in seven, so a horizon landing on a Saturday resolves on the next session rather than reading as "no data" and drifting into a void charged to your record. The window is fixed and the choice inside it is "the earliest session" — never "whichever price suits", and a third party can replay it exactly.

Why this matters: because the source and the entry price are locked in before the outcome is known, nobody — including us — can pick a friendlier price after the fact. The yardstick was chosen at the call, it's part of the immutable seal (see Sealed in time), and when the time frame is reached the resolution reads from that same source. A sealed source is never reinterpreted.

This is also why coverage grows carefully. Adding a topic means committing to a resolution source we can defend the same way — public, objective, and fixable in advance — for every asset in it.

Coming soon: the other topics

The other topics you'll see referenced around the site — AI stocks, tech, emerging markets — are coming soon, not silently absent. They already exist as axes the leaderboard is built to slice on ("who's best on AI stocks?"), but you can't yet publish scored predictions on them as their own topic. Note that many of those names are already predictable today under Equities: what's missing is not the price source, it's the classification — deciding which lines belong to "AI" is a judgement call, and one we'd rather make deliberately than by keyword.

We'd rather show a topic as coming than open one we can't resolve honestly. When a new one goes live, it arrives with its own named, sealed price source — the same discipline, extended. Equities are the first proof of that: they landed with their own frozen source, their own market-closure rule and their own stated limits, and they did not touch a single crypto prediction already sealed.

Next: Publish your first prediction · How a prediction resolves.