On November 10, the FAA announced that, following the Washington test, the agency "is now using SMART in the airspace of all 22 centers" across the country. The stated goals: fewer delays, less fuel burn, a better passenger experience. The developer is Boston-based startup Air Space Intelligence; the FAA awarded the contract, worth up to $875.9 million, in June. The Hacker News thread is mostly arguing about transparency.
The public materials are thin on mechanics. The FAA talks about "reducing delays and fuel consumption"; ASI talks about integration with the FAA and ATC. What the algorithm actually does — whether it redistributes traffic flows, suggests routes to controllers, or just runs in the background — doesn't follow from the quotes. On social media, the company denies any link to the ground stop: "The current stop is not based on SMART and is not the result of the system, any ASI product, or our integration with the FAA and ATC" (quote via NYT). Meanwhile, the FAA's nationwide flight suspension was being discussed too — it was lifted after 4.5 hours.
Our read: this story isn't about "AI replacing controllers." It's about how fast the government is pushing an experimental system into critical infrastructure. The FAA started a 90-day evaluation in Washington late last month, and a couple of weeks later it announced a rollout to all 22 centers. Formally, that's "after the test" — in practice, it's before the evaluation finished and before anyone saw its numbers.
Why it matters. For the market, it's a signal: government AI contracts are no longer about pilots and slide decks — they're about real control rooms. For passengers, it's a trust question: if the system gets something wrong, that's not a "bad search result" — it's delays and safety. And for the wallet: $875.9 million is going to a startup whose work in Washington hasn't been publicly measured yet.
We think the biggest risk here isn't technical — it's procedural. When an evaluation is still running but the expansion has already been announced, the regulator loses any incentive to publish inconvenient metrics. ASI wins: contract and attention. The losers are the people who asked for a pause — their opacity argument sounds stronger than the "everything is under control" response. If SMART really had nothing to do with the ground stop, the FAA could prove it — with logs, a report, a timeline. Until then, we're being asked to take it on faith.
There are plenty of weak spots. Everything here is assembled from a Hacker News thread: the NYT quote is a commenter's paraphrase, the contract figure and date come from a paraphrase of Politico, and the Gizmodo link is from the thread too. The link between the ground stop and SMART hasn't been officially confirmed, and the company denies it — but we haven't seen an independent verification. It's also unclear what "using it in all 22 centers" actually means: full operational mode, controller suggestions, or a test loop. Another commenter read Politico as saying the FAA wants an equity stake in ASI — that's an interpretation, not a confirmed fact. The primary sources worth checking: the FAA press release, the contract itself, and the report from the 90-day evaluation.
Forecast: within the quarter, the FAA will publish an interim report with numbers from the 90-day SMART evaluation. If no report appears, the expansion to 22 centers becomes a political problem, not a technical one.
Disclaimer: not officially confirmed — details are assembled from a Hacker News discussion and paraphrases of NYT, Politico, and Gizmodo coverage; ASI denies any connection between SMART and the ground stop.
Source: Hacker News
Source: news.ycombinator.com · post in Telegram