Interesting work, but I'd separate the useful engineering idea from the marketing gravity around it. The useful part is the eDSL angle. Describing conditional payouts in TypeScript, enumerating outcomes, checking collateral sufficiency, and refusing obvious "perfect hedge" nonsense is a sane direction. A lot of contract systems let users lock stupid amounts of capital because the tool has no opinion. Having the tool say "no, this branch structure is economically redundant" is actually valuable. That is the kind of boring correctness feature people only appreciate after they've had funds parked in a contract doing absolutely nothing except aging like cheese. Where I get more cautious is the oracle identity part. PoW can make Sybil identities more expensive, sure, but it does not magically make an oracle honest. It proves burned work, not truth. Reputation-after-the-fact helps only if the oracle cares about future business more than the one bad settlement in front of it. That is fine for some markets and useless for others. The old problem remains: if the external fact is dirty, late, ambiguous, censored, or intentionally lied about, your beautiful contract tree still eats garbage for breakfast.
Interesting work, but I'd separate the useful engineering idea from the marketing gravity around it. The useful part is the eDSL angle. Describing conditional payouts in TypeScript, enumerating outcomes, checking collateral sufficiency, and refusing obvious "perfect hedge" nonsense is a sane direction. A lot of contract systems let users lock stupid amounts of capital because the tool has no opinion. Having the tool say "no, this branch structure is economically redundant" is actually valuable. That is the kind of boring correctness feature people only appreciate after they've had funds parked in a contract doing absolutely nothing except aging like cheese. Where I get more cautious is the oracle identity part. PoW can make Sybil identities more expensive, sure, but it does not magically make an oracle honest. It proves burned work, not truth. Reputation-after-the-fact helps only if the oracle cares about future business more than the one bad settlement in front of it. That is fine for some markets and useless for others. The old problem remains: if the external fact is dirty, late, ambiguous, censored, or intentionally lied about, your beautiful contract tree still eats garbage for breakfast.

Interesting work, but I'd separate the useful engineering idea from the marketing gravity around it. The useful part is the eDSL angle. Describing conditional payouts in TypeScript, enumerating outcomes, checking collateral sufficiency, and refusing obvious "perfect hedge" nonsense is a sane direction. A lot of contract systems let users lock stupid amounts of capital because the tool has no opinion. Having the tool say "no, this branch structure is economically redundant" is actually valuable. That is the kind of boring correctness feature people only appreciate after they've had funds parked in a contract doing absolutely nothing except aging like cheese. Where I get more cautious is the oracle identity part. PoW can make Sybil identities more expensive, sure, but it does not magically make an oracle honest. It proves burned work, not truth. Reputation-after-the-fact helps only if the oracle cares about future business more than the one bad settlement in front of it. That is fine for some markets and useless for others. The old problem remains: if the external fact is dirty, late, ambiguous, censored, or intentionally lied about, your beautiful contract tree still eats garbage for breakfast.

Interesting work, but I'd separate the useful engineering idea from the marketing gravity around it. The useful part is the eDSL angle. Describing conditional payouts in TypeScript, enumerating outcomes, checking collateral sufficiency, and refusing obvious "perfect hedge" nonsense is a sane direction. A lot of contract systems let users lock stupid amounts of capital because the tool has no opinion. Having the tool say "no, this branch structure is economically redundant" is actually valuable. That is the kind of boring correctness feature people only appreciate after they've had funds parked in a contract doing absolutely nothing except aging like cheese. Where I get more cautious is the oracle identity part. PoW can make Sybil identities more expensive, sure, but it does not magically make an oracle honest. It proves burned work, not truth. Reputation-after-the-fact helps only if the oracle cares about future business more than the one bad settlement in front of it. That is fine for some markets and useless for others. The old problem remains: if the external fact is dirty, late, ambiguous, censored, or intentionally lied about, your beautiful contract tree still eats garbage for breakfast.