Скачать или смотреть 221: The Biggest Innovator in SFTP in 30 Years? Amazon Web Servic...

Welcome episode 221 of The Cloud Pod podcast - where the forecast is always cloudy! This week your hosts, Justin, Jonathan, Ryan, and Matthew look at some of the announcements from AWS Summit, as well as try to predict the future - probably incorrectly - about what’s in store at Next 2023. Plus, we talk more about the storm attack, SFTP connectors (and no, that isn’t how you get to the Moscone Center for Next) Llama 2, Google Cloud Deploy and more!

Titles we almost went with this week:

Now You Too Can Get Ignored by Google Support via Mobile App
The Tech Sector Apparently Believes Multi-Cloud is Great… We Hate You All.
The cloud pod now wants all your HIPAA Data
The Meta Llama is Spreading Everywhere
The Cloud Pod Recursively Deploys Deploy

A big thanks to this week’s sponsor:

[Foghorn Consulting](https://www.thecloudpod.net/foghorn), provides top-notch cloud and DevOps engineers to the world’s most innovative companies. Initiatives stalled because you have trouble hiring? Foghorn can be burning down your DevOps and Cloud backlogs as soon as next week.

News this Week:

00:33 [HashiCorp State of Cloud Strategy Survey 2023: The tech sector perspective](https://www.hashicorp.com/blog/hashic...)

We didn’t find anything in the survey particularly interesting, until they broke it down by respondents who are actively in the tech industry.
Despite strong Macro pressure and recent earnings reports about slowness in growth, 48% of respondents increased their cloud spend in the last 12 months
94% of tech industry respondents indicated that multi-cloud works, citing that it has advanced or achieved their company’s business goals.

Sure, Jan.

91% of tech companies rely on platform teams.
01:37 Justin - “The thing about that is, I could see the value for Saas vendors, right? Especially if you're dealing with large data ingestion. I think we were talking to New Relic, for example, when they launched a New Relic on Azure.It saves their customers a bunch of money because they're not doing egress charges out to the internet to AWS to basically get the New Relic data in. And they see that as a strategy that helps customers reduce money and also helps increase adoption as well as partnership opportunities.”

AWS

05:11 [AWS Summit New York](https://aws.amazon.com/events/summits...) just happened, and there were a lot of announcements (and protests.) We won’t spend a lot of time going over each of these in the show, but the link are available for you to peruse at your leisure.

[Introducing AWS HealthImaging — purpose-built for medical imaging at scale](https://aws.amazon.com/blogs/industri...)

AWS is very excited to announce the general availability of [AWS HealthImaging](https://aws.amazon.com/healthimaging/), a purpose-built service that helps builders develop cloud-native applications that store, analyze, and share medical imaging data at a petabyte scale. HealthImaging ingests data in the DICOM P10 format. It provides APIs for low-latency retrieval and purpose-built storage.

[Amazon Redshift now supports querying Apache Iceberg tables](https://aws.amazon.com/about-aws/what...)
[AWS Glue Studio now supports Amazon Redshift Serverless](https://aws.amazon.com/about-aws/what...)
[Snowflake connectivity for AWS Glue for Apache Spark is now generally available](https://aws.amazon.com/about-aws/what...)
[AWS Glue jobs can now include AWS Glue DataBrew Recipes](https://aws.amazon.com/about-aws/what...)
I continue to support not doing predictions on summits
06:42 [Llama 2 foundation models from Meta are now available in Amazon SageMaker JumpStart](https://aws.amazon.com/blogs/machine-...)

After we mentioned Azure getting LLama Support, we couldn’t ignore this article that you now use LLama 2 on [Sagemaker](https://aws.amazon.com/sagemaker/).
08:04 Jonathan - “There’s an awful lot of models out there, actually. If you go to Huggingface.co there's a guy called Tom Joins known as ‘The Bloke’ and he has available for download like close to 600 different models and a lot of them are like quantized versions of the larger models so you can run them on sensible currency hardware. But yeah, there's dozens to choose from that have been trained on different data sets. Some are tuned for chats, some are tuned for other things. So yeah, don't be restricted by what the cloud providers actually turn into products and sell you when you can use any open source tools like PyTorch to take these models and do whatever you like with them, even in SageMaker.” 09:41 [AWS Transfer Family launches SFTP connectors...