BigQuery

right um in our the highest and most valuable use of telium is not that we have to hold the audiences is that we glue together on a trusted fabric all of these systems you know and we do it in a very in a very trusted privacy compliant way and we do it with high performance around the globe and with effectively zero data loss you know enterprise grade reliability

Yeah, and so how do you activate data that lives in Snowflake or Databricks or BigQuery, Redshift, or one of these in real time without creating a shadow copy or a second source of

truth.

We have a capability called Cloud Stream and we'll go extract an audience out of the CDW and activate it in real time out to whatever activation point they want to go to.

And I honestly didn't want to leave that role at Tasktop.

We were doing amazing things and working with Mick and the team was fantastic.

But throughout my, I guess, last 20 years, maybe 15, Ken Schwaber had been increasingly a mentor to my, to my work.

And he's a very interesting mentor.

And I think over the last 11 years, we haven't necessarily solved some of those problems, but we have definitely made them more transparent in organizations.

We have really leaned into scrum about value scrum you know the work i did with lean ux bringing that in the work i did in kanban to bring flow in the work i did with nexus to talk about scaling the work recently we've been doing a lot about the agile product operating model and now obviously how ai is influencing that and how You know, the secret source for Scrum and Agile isn't anything particularly secret.

It is aligned to a clear problem domain, empower teams to make decisions and provide support to get rid of the things that get in the way of them actually delivering value to customers.

It's really, really simple.

A simple automation.

You can configure it to send to the BigQuery table.

So you have all your metrics data.

And once the data is available over there, you can do the automation in terms of the dashboarding.

When unbacked balances reach real reserves.

Next, Initia, BigQuery's original investigation traces unbacked balances being redeemed through bridges on October 23rd.

It estimates about $2.73 million paid out from reserves before the network halt.

Think of the security question this way.

Circle CCTP transfers native USDC through burning and minting.

We should understand the actual route rather than assume every bridge works alike.

So what failed? BigQuery extinguishes the face value of fabricated balances from the real assets withdrawn.

Its tracing supports a reserve loss story.

Turns out each of the big hyperscalers has their own data warehousing platform.

Google's got BigQuery.

Azure's got Synapse.

AWS has Redshift.

So there have been like a couple Spark connectors that have already been implemented as well.

There's also vendor connectors like Databricks has one, for instance, BigQuery has one, which ADBC already works on and standardizes.

But the benchmarks on this early prototype show, and this is highly variable, so, you know, keep a... a grain of salt in terms of the final figure here, but at least like a 3 to 10 times read throughput compared to JDBC on column-oriented databases, which we have been talking a lot about columns and the benefits of columns and in-place column updates, things like that.

So columns for the win, I guess, and faster columns.