We recently scaled Shield, @stripe's transaction fraud model, and improved recall by more than 8%. A few lessons that apply well beyond fraud:
4
3
53
5,610
We see you—and your agents—building on Stripe data.
Now you can sync your Stripe data to Postgres in real time. No code, no webhooks to maintain, and free to get started.
Hello, real-time data. Show us what you build 👀
stripe.com/blog/real-time-st…
14
12
276
40,652
10 years of @StripeJapan 🇯🇵
I celebrated in Tokyo with @mshrnakagawa on @tbs_bloomberg, talking about trust in AI-powered payments, agents that transact, and how AI is changing work.
Also best meals of my life 🍣
「人間ではなくAIが買い、AI同士が取引する」
OpenAI・Anthropicも頼る米決済サービス「Stripe」のデータから見えてくる【AIエージェント経済学】を読み解きます
Stripeのデータ・AI責任者であり、ハーバード大学で博士号を取得した経済学者のエミリー・サンズさん @emilygsands に、MCの中川雅博 @mshrnakagawaが迫ります
2
5
11
2,407
Is fine-tuning LLMs worth it? Generally yes when you need precise evaluations plus the flexibility to use rich, heterogeneous evidence.
Take merchant fraud reviews at @stripe. We fine-tuned an open-weight 4B model on review labels and read a fraud score from its yes/no token probabilities. Average precision rose 24% over prompting, and the model beat human reviewers at the same recall.
Linear probes suggest an important driver is that fine-tuning made *legitimate* businesses' signals easier to decode.
Full post: stripe.dev/blog/training-llm…
2
4
18
1,211
I sat down with @natebjones to discuss how @Stripe is building economic infra for AI. We cover:
-Trust in the agentic economy
-Why @patrickc refers to 1/1/26 as day one of the singularity
-Building with @cursor_ai to stop fraudsters from stealing tokens
piped.video/YTG0rdHPTDE?si=KdPF…
8
10
64
10,302
At the speed with which AI can change the economy, quarterly reports are too laggy. The @stripe SaaS Index is weekly, same-business pay-in volume across ~72k SaaS companies--so it moves right alongside demand.
And here's what it's telling us: SaaS growth accelerated straight through the SaaSpocalypse selloff. Yes, accelerated. So more like a RenaiSaaS.
Our SaaS Index is a weekly measure of non-AI SaaS payments processed on Stripe. We constructed it to be robust to Stripe's own business mix: it's compositionally-adjusted over time & filters out extreme ramp-ups/downs. We plan to update it weekly here. /2 stripeeconomics.com/p/stripe…
1
5
24
2,909
The hard part of enterprise AI has little to do with the models; it's permissions, context, and governance.
Sharadh on how a two-week v1 from 1.5 engineers became the thing 10,000+ Stripes now use every week. (And how years of boring platform work is what made it possible.)
If you're interested in solving hard, technical problems at scale, we're hiring! stripe.com/careers
The team at @stripe is setting the standard for internal AI platforms: minion coding agents, a custom prototyping rig, and now their company brain, Kai.
On today's episode of How I AI, Sharadh shows us how 1.5 engineers and 2 weeks got them a company brain, including:
- projects as governance
- skill routing + telemetry
- a skills platform that works for 10k teammates
Plus, he and I debate the merits of gentle parenting your AI (esp when your company is running evals.)
Full episode on YT: piped.video/watch?v=AbZODZ_4…
2
3
12
2,788
Emily Sands retweeted
New from @llamamrama on Stripe Economics on the geography of new business formation in the AI era.
A plausible hypothesis was that AI-era businesses would value agglomeration effects more. But so far, post-2024 new businesses are even more disbursed.
stripeeconomics.com/p/ai-and…
1
6
20
6,873
