David Yaffe (@dyaffe) profile picture

Founder of https://teapaw.org/t.co/bIcz7CbBby | Previously founded Arbor (acq @LiveRamp) & Product @invitemedia (acq @Google)

New York
Joined May 2009
Who would have guessed that having access to unimaginable intelligence through AI makes great people work more?
1
24
This Friday, the Estuary and CorrDyn teams walk through our warehouse benchmark live. 244 queries, 7 warehouses, 1 GB to 1 TB. Bring questions, especially the curveballs. go.estuary.dev/dwh-live
2
86
The HubSpot materialization works with custom objects as well now. When we first put it out in September it only did the standard ones, but as it turns out, most of the useful data lives in custom objects. Continuously hydrating a CRM is new territory for us but it's exciting to see what people are building with this capability.
1
28
DuckDB can now plug straight into your Postgres or MySQL now and run queries against it - pretty nice for quick ad hoc stuff. But think about who’s asking. That query actually runs on your live production database, and the thing most likely to fire off some giant query at a random, inconvenient time is an AI agent. I’d at least point it at a copy of the database. Better yet, a copy that’s already kept up to date somewhere else… that’s the version I’d let an agent anywhere near.
3
2
90
It's great to see that interest in our startup program has exploded in the last few weeks. My guess as to why is that early teams are deciding they need real-time data sooner than they used to, especially to power their agentic workflows. If you're interested: early stage startups get $2k of platform usage for free.
3
53
Keep using agents in the warehouse for analytics. Acting the moment something happens is a different job, and a scheduled run against a warehouse isn't how I'd do it. We're working on running that part in the pipeline, on the same data that keeps the warehouse current.
19
Obviously hilarious goading but he's not wrong. I'd much prefer to bring my own model to a data system and not be stuck with what they're offering.
Replying to @RamaswmySridhar
I don't think we should fight. The two companies are complementary. Since you're not doing this AI stuff, I really think Snowflake should consider getting its AI tokens through Databricks. You can get both cheap OSS models as well as frontier (OAI/Ant/Gemini/Grok) models from our platform. It all goes through Unity Gateway which has Smart Routing, Budgets, and Access Control. You can even use Omnigent with any harness, including CoCo. Most of our customers used this to reduce token spend ~30% last few months while token count went up. Genie lets you then track, control, manage that spend, all your financials, marketing, etc. Quick POC?
3
94
There are more cloud data warehouses than ever, and they keep converging. Most of the features that used to separate them are table stakes now, which leaves performance and cost as the real difference, and I think buyers are watching both more closely than they ever have. The trouble is that almost every published benchmark comes from a vendor, and it tends to favor that vendor. So we ran our own on seven major warehouses: @Snowflake, @databricks, #BigQuery, #Redshift, #MicrosoftFabric, @ClickHouseDB and @motherduck 244 queries, data from 1 GB to 1 TB, and up to 100 people querying at once, at every compute size each vendor sells. Every engine got identical data, loaded through one Estuary capture. I don't think there's a single best warehouse, and the report says so… the right one depends on your workload, how many people use it at the same time, and what you're willing to pay per hour. Doing this research is time and money consuming. After months of work, we decided to make the whole report (and data behind it) free, to help you make better buying decisions. The Complete 2026 Cloud Data Warehouse Benchmark is available here: go.estuary.dev/dwh
1
7
92
Open source used to be a funnel. You self-host, hit the hard parts (upgrades, scaling, backfills, the 3am page), and pay for the managed version. With AI the hard parts are one prompt away from a working answer. Give away enough and there's very little left to pay for.
2
38
We're letting people write their own connectors on Estuary. You write the Python, we run it, and the data lands in the same collections as every other source. Looking for a few teams with a source we don't cover (internal API, weird SaaS, legacy system) to try it first.
33
Cloudflare shipped a model that reads an email or an image and returns typed values, forecasts and confidence scores in one pass, no tokens generated. Cool. It looks like the model is becoming the easy part. Getting every email, Slack message and ticket to it, in order, with nothing dropped, is still on the pipeline.
51
Super excited to announce @deel is an @EstuaryDev customer. Integration costs are down 50% and they have access to the freshest data. Exactly the results we love to see 👏 Case study here: estuary.dev/success-stories/…
1
33
Canceled our contract with our BI tool this week. Would have probably done it anyways but @claudeai dashboards made it a no brainer.
5
87
Good piece, pretty close to where we’re going, but the key missing piece is mediated execution. Bring the data to the agent, don’t let it reach into prod.
The Modern Data Stack is over, long live the Post-AI Data Stack🫡🤖 I wrote about how technological shifts change what a data team can build, the features of post-AI data stacks, what we've built at @tryramp, and what data teams can learn from @nbcsnl. iandmacomber.com/blog/post-a…
73
David Yaffe retweeted
Another migration from Go -> Rust. Reading about how Estuary rebuilt the engine behind their CDC platform. I thought it was interesting how the lack of a garbage collector was a plus and how saving staging on disk was preferred over in-memory. estuary.dev/blog/why-we-rebu…
3
77
6,182
Too many teams shipped their AI strategy by putting ‘agent’ in the h1 and changing nothing else. If agents actually changed your product, your product would have changed.
3
73
text2sql is a red herring The industry spent many years shoehorning SQL into places it doesn't belong with the only real argument that people already knew SQL. Agents don’t need the familiar interface.
2
66
it's honestly that simple.
how to make your site readable by ai agents: 1. add an llms.txt so search tools stop guessing urls 2. check your server logs, not your analytics 3. serve a markdown version of every page 4. support content negotiation headers 5. rethink your JS-only widgets
1
64
Agent plugins are foundational for a new app store. Who's optimizing for it? We’re in 100 coding assistants and I have almost no idea how an agent decides to pick us over something else. Seems like a bigger distribution question than SEO ever was.
1
2
56