Topic
38 facts from 12 episodes, in date order. These are what speakers said on the shows, each linked to the moment it was said. They aren't independently verified.
2017-2020
Kenny Song's Google Brain team's infrastructure work included TensorFlow, TensorFlow Extended (TFX) and parts of Google Cloud's AI platforms.
“this included TensorFlow and also other platforms like TFX, TensorFlow Extended, and some of the work on Google Cloud's AI platforms and so on.”
2018
Google launched Kubernetes in 2018 as a multi-cloud strategy for its third-place cloud business.
“2018 is their first very clever strategic decision. They launched Kubernetes.”
2020s
Ramin Mohammadi says Google Cloud's Rules of MLOps laid out that successful machine learning needs real engineering practices such as containerization, CI/CD automation and monitoring.
“Some resources like, uh, Google Cloud, Rules of ML Ops, you know, it. It re laid out these new realities that successful ML needs a whole suit of real engineering escapes.”
2022
Rob Enslin was a Google Cloud executive whom Daniel Dines brought into UiPath to help steer the company after its public listing.
“Rob Enslin, a Google Cloud executive he brought on to help steer UiPath after it went public.”
2023
Google Cloud turned profitable in 2023 and has an annual revenue run rate above $50 billion growing about 30% a year.
“They also flipped to profitability in 2023. And today, they're over $50 billion in annual revenue run rate. It's growing 30% year over year.”
2023-2026
Matt Renner says most early enterprise AI proofs of concept failed, defined as not ending up in a production system that drove value.
“the majority did fail, and failure being defined as it didn't end up in a production system that was driving value, right?”
2025
In the UK telecom app, customers point their phone camera at home equipment so agents can guide troubleshooting.
“They will have the individual, um, point their camera at the equipment that they have in the house.”
2025
In a home-design visualization project, early AI outputs sometimes placed couches on their sides, requiring explicit physics instructions for the evaluating agent.
“A funny little quirk is one thing that we found when we were first starting to do this is that, you know, manipulating objects in three-dimensional space. Sometimes we'd have like a couch sitting on its side”
2025
The home-design agent pipeline added a second evaluation that checks whether a suggested item is in the retailer's inventory.
“is this item actually, you know, an item that's in our inventory, you know, a very specific item. And so that was a second evaluation and critique.”
2025
Google BigQuery includes native AI integrations and inference within the data system.
“today in BigQuery in Google Cloud, we have native AI integrations and native AI inference in the data system itself.”
2025
A bank working with Google sped up its research function using agents, cutting turnaround from days to hours or minutes.
“we're working with a bank right now that has seen the ab- that has sped up, uh, their research function.”
2026
Sarah Friar said OpenAI announced it was live on Google Cloud Platform, so its products are available on Amazon, Azure, Google Cloud and Oracle.
“we just announced this morning we're live on GCP, on Google, so we now can meet our customers on every cloud that exists”
Undated
Google Cloud's Vertex AI Model Garden includes domain-specific models such as Med-PaLM and Sec-PaLM.
“That's where we have domain specific models, so it's Med Palm, uh, sec palm.”
Undated
Kawal Gandhi said open source models such as Llama and Stable Diffusion were coming to Google Cloud's platform.
“I'm really excited to see, like Llama, stable diffusion, uh, other variations coming onto the platform”
Undated
Gandhi said Google Cloud has invested in model operations, including handling model drift, tooling and safety.
“Because we've spent a lot of time really specializing around the model drift, operations, tooling, safety around it.”
Undated
Gandhi said fine-tuned adapter models and related artifacts stay within the customer's own tenant on Google Cloud.
“The adapter model, the certification, everything stays in their tenant.”
Undated
Gandhi said GPUs are widely available and many engineers are trained on them, which is a consideration when choosing between TPUs and GPUs.
“Because there is this that was, you know, externally available, and a lot of people are trained up.”
Undated
Gandhi said the NVIDIA GPU shortage comes up in customer conversations with Google Cloud.
“It definitely is mentioned, uh, as we think about it in the conversations, uh, it's always the the discussion that comes up.”
Undated
Gandhi said Google Cloud aims to make new models such as Llama available in its Model Garden so customers can use them within existing applications.
“can we make it available like Llama inside our model garden? And then they can use it inside their application”
Undated
Gandhi said Google Cloud is exploring synthetic data, including for use in regulated industries.
“And we're also looking at synthetic data as well. Can we use in regulated industry some of the synthetic stuff”
Undated
Google Cloud offers data labeling and related services through a partner marketplace integrated into its platform.
“So we have, uh, the partner ecosystem through that, we provide a lot of partners, so exactly the marketplace.”
Undated
Gandhi said data from partner labeling services is kept in the customer's project and not shared with other customers.
“So we make sure those pipelines and entitlements stay in the project.”
Undated
Weights & Biases used MySQL because Google Cloud offered managed MySQL but not managed Postgres at the time.
“We used MySQL as the like data store for for everything because at the time we were on Google Cloud. And all they had was a managed MySQL offering, they didn't have managed Postgres.”
Undated
Bret Taylor says cloud computing consolidated into a small handful of infrastructure-as-a-service providers, such as Amazon Web Services, Azure and Google Cloud, that account for most cloud capital spending.
“you ended up with a small handful of infrastructure as a service providers that represent the vast majority of the capex investment in cloud”
Undated
GNP deploys its systems mainly on Google Cloud.
“we deploy our systems mainly in Google Cloud”
Undated
Citibank's first AI use cases with Google Cloud focused on efficiency and compliance, such as making large volumes of documents easier to search.
“So I can use Citibank as a good example. So when we started working with Citibank and AI, their first few use cases were exactly down the path you just mentioned”
Undated
Google Cloud and Citibank built a system called CitiSky, which Matt Renner describes as focused on customer experience.
“So you look at the system we built called City Sky with them, and and again, in this specific case with that product, uh, it's all around customer experience.”
Undated
Honeywell used its manufacturing and product data to create a new AI-based services business that generates revenue.
“What if we were to actually create businesses with AI and actually create a revenue stream around a services business that was related to a lot of these products”
Undated
Matt Renner says Google Cloud spent about fifteen years convincing customers to move data to the cloud.
“we spent, well, in in in Google Cloud's case, about fifteen years convincing people to move data into the cloud, but not just move it.”
Undated
Matt Renner says many early cloud adopters used lift and shift, moving data without modernizing it, which he likens to outsourcing.
“There's a fair number of folks who just moved it over, lift and shift and didn't really modernize where the data was. It was just basically an outsourcing scenario.”
Undated
Google Cloud offers BigQuery as a data store and is developing an agentic data lake, according to Matt Renner.
“We have a phenomenal store in BigQuery, and now we've got our our, uh, a lot of things going on with with our our agentic data lake that we're putting out there at this point.”
Undated
Matt Renner says Gemini Enterprise, Google Cloud's agent orchestration platform, is the fastest-growing product in the company's history.
“we have one called Gemini Enterprise, which is the fastest growing product in our history to ever hit the ground.”
Undated
Matt Renner says the A2A protocol is currently the most dominant standard for letting agents from different vendors talk to each other.
“A two A is the is the most dominant one at this point. So it can talk to agents, uh, in your orchestration platform”
Undated
Matt Renner describes Google Cloud's agent orchestration platform as a neutral 'Switzerland' that works across vendors' agents.
“And and again, we're we're more of a Switzerland, i if you will, orchestration platform for agents as you go through it.”
Undated
Matt Renner says Google acquired the cybersecurity consulting firm Mandiant about seven years before the interview.
“We made an acquisition of Mandiant about seven years ago as we were starting entering in into this era”
Undated
Matt Renner argues that AI models mainly find vulnerabilities, so customers also need tools to remediate, track and monitor them.
“the model's only going to scan for the most part. It's going to find your vulnerabilities. But now what?”
Undated
Matt Renner predicts most customers will use multiple AI models to scan for security problems.
“most customers, I think, are going to use multiple models before it's all said and done to make sure they've found all the problems”
Undated
MongoDB's managed service, Atlas, can be deployed in the data centers of Amazon, Google and Azure.
“which you can deploy your instances on any cloud hyperscaler data center you'd like amongst Amazon, Google, and Azure”
Episodes