ericdodds

@RhysSullivan I go to goodwill to run open cloth

cloth code

@rauchg I stopped using AI to write posts on my personal blog for this exact reason. What’s really interesting is that working those muscles seems to make me better at using AI to write.

If you don’t use your body, it atrophies, so you go to the gym or on a run. If you don’t use your brain, it atrophies, so you… What’s your plan?

@Adiiizi @vercel Great. Thanks for following up.

@ericdodds @vercel Working perfectly now.

Building is no longer the bottleneck. But that doesn't mean your agents are secure, reliable, or easy to audit and maintain in production. Vercel's self-driving infrastructure fixes this. vercel.com/blog/anyone-ca…

@Adiiizi @vercel Thanks for the heads up. We fixed the link this morning and I just double checked it. Could you refresh and check again?

@vercel Hey @vercel @ericdodds, I think redirect link for Sandboxes is probably wrong in the blog. It's: vercel.com/docs/vercel-sa…

@dfeinition It's incredible to have a front seat on this ride.

Progress of AI. March '23 → Jan '26

@aryamankhawow @v0 Yes, a debugging mode would be great - there’s a lot that could be done there as far as automatic checks, surfacing errors, restarting the dev server when appropriate, etc.

@ericdodds @v0 Would love chat links where v0 didn’t automatically read logs! It does have the ability to do so, so we may need to prompt tweak a bit there. “Debugging mode” is super interesting tho

Sometimes, you want @v0 to understand your intent better. "I just want a prototype, don't write API routes" "Answer my question, don't build an app" We're working on something to make this magical. What other intents would you like v0 to capture along with your prompt?

@dfeinition When your vibes are right, they're right

not ashamed of this^

The moment data enters the conversation, my contributions drop to zero

"If we don't love it, we don't even bother."

@ankrgyl @RudderStack @fouadmatin @RillData @duckdb @posthog Also, we're actually working on streaming Vercel logs into RudderStack so that we can analyze that data in a number of ways (product analytics tools, visualization on @SnowflakeDB , etc. Pretty neat stuff!

@RudderStack @fouadmatin @RillData @duckdb @posthog I tried Rudderstack first actually, and churned because the debugger was too slow (compared to Segment’s)

What's the best analytics system for next.js/vercel (e.g. is segment good, or something else)? An important requirement: must be able to write raw logs to S3.

@ankrgyl @RudderStack @fouadmatin @RillData @duckdb @posthog We could definitely do a better job of making that explicit to new users. I'd be happy to set you up with an account where you can test latency and scale at an enterprise level.

@ankrgyl @RudderStack @fouadmatin @RillData @duckdb @posthog Thanks for the feedback, @ankrgyl. Our architecture is warehouse-native, meaning we don't store events, so we have to explicitly open a connection between the data plane and control plane to stream live events. On the free tier, this can take a few seconds.

Transformations was a huge reason I transitioned to RudderStack from Segment in my consulting days. It's one of those features that so useful it's hard to remember running pipelines without it.

@barrald @_hex_tech This might help you interpret the poll results:

Should we enable pie charts in @_hex_tech chart cells? We are undecided after several internal ballots so I turn this decision over to the masses

This is one of the top conversations I've had in the past year—about data, starting things, the deeps of technology and, most importantly, human empathy. A true pleasure, @frankmcsherry and @narayanarjun

@tersmantoll @cbrodzky Wow, that is heavy duty. Are you running at extreme volume?

@ericdodds @cbrodzky Ok, so here is the correct description: All actions from Matomos Visitor logs/raw data gets imported into Apache Druid as a data lake with the help of SQL. And Superset then - also with SQL - brings data from Druid to the report UI. Never to late to correct one self :)

ICYMI: About a month ago, Google began enforcing a quota limit in Looker Studio for the GA4 connector. This means that if you reach your limit, which most of us prob will, your dashboard will look like this. There are a few solutions - none of which I like at this time. 😡

@tersmantoll @cbrodzky Wow, very cool! Sure, would love to chat to learn more about what led to this kind of custom build. Feel free to send me a DM.

@ericdodds @cbrodzky If you want that even more technical i'll have to ask my colleagues who manages the kubernetes setup for our Matomo Cloud, and Apache Superset respectively and get back to you :) Or we could talk live

@iamphilwade @cbrodzky @RudderStack @mixpanel 100%, that's the big risk. I really hope they don't because I think this stack is something that can really help smaller companies punch far above their weight class in terms of data. It's Google, though, so I'm not holding my breath :)

@ericdodds @cbrodzky @RudderStack @mixpanel Once they remove or lower the free tier the cost might become meaningful. Let's hope they don't :)

@iamphilwade @cbrodzky @RudderStack @mixpanel Philosophically I think this is a good thing—it makes advanced data projects much more accessible to many more companies. But at scale the infrastructure ultimately decreases option value and flexibility for data teams compared to other architectures.

@iamphilwade @cbrodzky @RudderStack @mixpanel At scale, the system is structured to generate revenue for Google Cloud at the data infrastructure level, which is a brilliant move by Google. At ~75% market share with Google Analytics, imagine the potential here.