VQV Signal
According to a post on X by @tanayj, Duolingo CEO Luis von Ahn said the company's AI video-call feature cost about $0.30 per call in inference at launch. The company has since reduced that to "well under $0.01 per call," mainly by moving to open-source or open-weight models. The post is a secondhand quote of the CEO, and the snippet does not give further details on the models used or the timing of the change.
According to a post on X by @tanayj, Duolingo CEO Luis von Ahn said the company's AI video-call feature cost about $0.30 per call in inference at launch. The company has since reduced that to "well under $0.01 per call," mainly by moving to open-source or open-weight models. The post is a secondhand quote of the CEO, and the snippet does not give further details on the models used or the timing of the change.
AI-assisted summary based on the listed source.
@tanayj (24 likes) quotes Duolingo CEO Luis von Ahn: the AI video-call feature cost about $0.30 per call in inference at launch, and the company has brought that to "well under $0.01 per call," mainly by moving to open-source or open-weight models. Why it matters: It shows how consumer apps can make heavy AI...
Builders of consumer AI features can see that switching away from frontier closed models may sharply lower inference costs for heavy features like live video. Product and finance teams at other apps may want to weigh similar model choices against quality and reliability. Watch whether Duolingo or its executives publish more detail on the models and performance tradeoffs.
VQV organizes public signals from inspectable sources. It does not independently verify the underlying report.
Signal Strength reflects source quality, relevance, freshness and evidence. Public Interest helps organize discovery; it is not proof of truth.
VQV surfaced this signal because it is recent, relevant to AI Video, connected to Scobleizer News Pulse (X).
No login, cookies, social SDKs, or automatic posting.
