March 13, 2025

295: Skype follows Chime to the Grave

1 hour 3 minutes

Welcome to episode 295 of The Cloud Pod – where the forecast is always cloudy!

Welp, it’s sayonara to Skype – and time to finally make the move to Teams. Hashi has officially moved to IBM, GPT 4.5 is out and people have…thoughts. Plus, Google has the career coach you need to make all your dreams come true.*

*Assuming those dreams are reasonable in a volatile economy.

Titles we almost went with this week:

Someday we’ll find it, the rainbow connection, the lovers, the cloud dreamers, and Me

Dreamer, you know you are a dreamer

You may say I’m a cloud dreamer, but I’m not the only one

May the skype shut down

Q can tell me that my python skills are bad

How many free code assistance does Ryan need to be a good developer: ALL OF THEM

Oops honey I spent 1M dollars on oracle

Latest Cloud Pod Reviews: “It’s a Lemon”

A big thanks to this week’s sponsor:

We’re sponsorless! Want to get your brand, company, or service in front of a very enthusiastic group of cloud news seekers? You’ve come to the right place! Send us an email or hit us up on our slack channel for more info.

General News

01:04 On May 5, Microsoft’s Skype will shut down for good

In what we swear is the 9th death for Skype, Microsoft has announced that after 21 years (with 13 of those years under MS Control,) Skype will be no more.

For real this time. Really.

May 5th is the official last day of Skype, and they’ve indicated you can continue your calls and chats in Teams.

Starting now, you should be able to use your Skype login to get into Teams.

For those of you who do this, you’ll see all your existing contacts and chats in Teams.

Alternatively, you can export your Skype data, specifically contacts, call history and chats.

Current subscribers to Skype Premium services will remain active until the end, but you will not be able to sign up for Skype at this time.

Skype dial pad credits will remain active in the web interface and inside Teams after May 5th so you can finish using those credits.

03:37 Matthew – “I think there’s a lot of people and, you know, at least people I know in other countries to still use Skype, like pretty heavily for like cross country communications, things along those lines. So I think a lot of that is that there probably is still a good amount of people using it. And this is just, Hey, they’re trying to make it nicely. So how, you know, nice and clean cut over for people versus, you know, the Apple method of it just doesn’t work anymore. Good luck.”

04:41 HashiCorp officially joins the IBM family

IBM has finished the acquisition of HashiCorp, which they had announced last year.

Armon Dadgar wrote a blog post reflecting on the journey that Hashicorp has been on; he talks about the future and that his goal is to have Hashicorp in every datacenter.

He says while they have made strides towards that goal, he feels incredibly optimistic with IBM, since they gain access to their global scale and increased R&D resources.

There are also integration opportunities of IBM and the RedHat Portfolio. Integrating Terraform for provisioning with Ansible for configuration management will enable an end to end approach to infrastructure automation as code, while integrating terraform with cloudability will provide native Finops capabilities to manage and optimize costs at scale.

Vault integration with OpenShift, Ansible and Guardium will bring world-class secrets management to those platforms and reduce the integration burden on end users.

05:44 Justin – “BM is gonna make a bunch of money if they force me to use Vault and Terraform Enterprise for all those capabilities. you know, HashiCorp was never shy to charge you at least $400,000. That was the starting price for pretty much everything.”

AI Is Going Great, Or How ML Makes Money

06:34 Introducing GPT-4.5

OpenAI has launched GPT 4.5, their largest and best model for chat yet.

GPT 4.5 is a step forward in scaling up pre-training and post-training.

Early testing shows that GTP 4.5 feels more natural. With a broader knowledge base, improved ability to follow user intent and greater “EQ” make it useful for tasks like improving writing, programming and solving practical problems.

They expect it to hallucinate less.

And on that note….

08:08 Hot take: GPT 4.5 is a nothing burger

Gary Marcus, author of rebooting AI, and founder and CEO of geometric intelligence (acquired by Uber) called Chat GPT 4.5 a nothing burger.

He previously predicted that GPT 4.5 wouldn’t be that impressive, and that the pure scaling of LLMs (adding more data and compute) has hit the wall.

He claims he was right. Hallucinations didn’t disappear, and nor did stupid errors.

He points out both Grok 3 and GPT4.5 didn’t fundamentally change anything, and both are barely better than Claude 3.5.

He quotes other AI forecasters who moved projections for AGI to later, and even pointed to Sam Atman’s rather tepid tweet regarding GPT 4.5.

Sam Altman also says they didn’t drop plus and pro at the same time because it’s a giant, expensive model and they need tens of thousands of GPU’s to roll it out to plus tier. He also says its not a reasoning model and it won’t crush benchmarks

09:13 Ryan – “It’s interesting because it’s in the consumer space, like you got to have flashy changes that dramatically change the user experience, right? So it’s like you always want to do incremental improvements. But if you’re announcing large bottle stuff, you know, it’s going to have a huge effect on your stock value. If the new stuff is just more expensive and more of the same. So it’ll be fun to see as they navigate this because it’s a new business model and uncharted territory.”

09:15 “It’s a lemon”—OpenAI’s largest AI model ever arrives to mixed reviews

The bad reviews for 4.5 weren’t just from Gary Marcus.

Ars Technica reported that it’s a “lemon”. Ouch.

Big, expensive and slow, providing only marginally better performance than GPT 4o at 30x the cost for input and 15x the cost for outputs.

10:16 Microsoft urges Trump to overhaul Biden’s last AI-chip export curbs

MSFT is urging the Trump Administration to ease export restrictions imposed on AI chips. Microsoft says the rules disadvantage allies, including India, Switzerland and Israel, and limit the ability for US tech companies to build and expand AI data centers in those countries.

Tighter US restrictions on the exports of advanced AI chips to Beijing are keeping American chipmakers and big tech from serving one of the largest markets for semiconductors, accelerating a global race for AI infrastructure dominance.

Microsoft says this will force some allies to turn to the Chinese market in the absence of sufficient supply of US tech. Left unchanged, the rule will give China strategic advantage in spreading over time its own AI technology, echoing its rapid ascent in 5G telecommunications a decade ago.

12:21 Ryan – “Which is basically what we saw with DeepSeek. They basically said, well, we can’t get these chips, so we’re going to figure out a cheaper way to build a model and then cause everyone to have pain. But the other reality is that I’m sure China is getting access to all these chips through some other country who doesn’t have quite the same restriction controls. They buy all the chips from the US, then they sell them on the dark market to China, I’m sure, if they really wanted them.”

AWS

13:16 AWS Chatbot is now named Amazon Q Developer

AWS Chatbot is now called Amazon Q Developer.

The new name recognizes the integration of Amazon Q Developer, the most capable generative AI powered assistant for software development, in Microsoft Teams and Slack to manage and optimize AWS resources.

With Q Developer, customers can monitor, operate and troubleshoot AWS resources in chat channels faster.

Customers can quickly retrieve telemetry and ask questions to understand the state of their resources.

14:03 Justin – “So AWS Chatbot is a very simple, I’m going to make a request and I have to use a certain syntax in the AWS chatbot to Slack. And then it calls the API and it returns data from the API that Amazon provides that I’ve synchronized and I have authorized. And it provides accurate data back to me. Amazon Q does not provide reliable data ever. It provides hallucinations. So if I ask it like how many Graviton based computers am I running in this region? And it comes back and says 32. Can I trust that there’s 32 boxes running or do I have to go double check it now because you’re using an LLM in the middle of this thing that doesn’t know what the hell it’s doing.”

21:06 Amazon ECS adds support for additional IAM condition keys

ECS has launched 8 new service-specific condition keys for IAM.

These new condition keys let you create IAM policies and SCPs to better enforce your organizational policies in containerized environments.

IAM condition keys allow you to author policies that enforce access control based on API request context.

With today’s release ECS has added condition keys that allow you to enforce policies related to resource configuration (ecs:task-cpu, ecs:task:memory and ecs:compute-compatibility), container privileges (ecs:privileged), network configuration (ecs:auto-assign-public-ip and ecs:subnet) and tag propagation (ecs:propagate tags and ecs:enable-ecs-managed-tags) for your applications deployed on ECS.

https://docs.aws.amazon.com/AmazonECS/latest/APIReference/API_CreateService.html

23:44 Matthew – “It’s a subset of the create service, which has grant permission to run and maintain the desired number of tasks from a specified task definition via service. So I think I might be right with the CPU task in there, where you could say you can’t create a CPU of a certain thing.”

26:55

Announcing extended support for Kubernetes versions for Amazon EKS Anywhere

AWS is announcing extended support for K8 versions of EKS Anywhere.

With extended support for K8 versions for EKS Anywhere, you continue to receive security patches for clusters on any K8 version for up to 26 months after the version is released in EKS anywhere.

Extended support for the K8 version for EKS anywhere is available for K8 1.28 and above.

27:20 Justin – “So, if you’re worried about the long-term supportability of Kubernetes and you don’t want to upgrade it every month, as you probably should, you can now get 26 months of support.”

27:55 Get insights from multimodal content with Amazon Bedrock Data Automation, now generally available

Announced at Re:Invent, Amazon Bedrock Automation is a feature to streamline the generation of valuable insights from unstructured, multi-modal content such as docs, images, audio and video.

The takeaway here is reducing the development time and effort to build intelligent document processing, media analysis, and other multimodal data-centric automation solutions.

Now, this capability is generally available with support for cross region inference endpoints to be available in more regions and seamlessly use compute across different locations.

Based on feedback during the previous, they have also improved accuracy and added support for logo recognition from images and videos.

GCP

29:24 Get coding help from Gemini Code Assist — now for free

Google is giving you Gemini Code Assist for individuals for free.

If you can’t sell it – giving it to engineers and then going after them for licensing violations is always a great move.

31:47 Discover Google Cloud careers and credentials in our new Career Dreamer

Google says if you have never worked in the cloud, it can be hard to know where to start.

Even if you’re a seasoned cloud architect, how do you pivot to your next big thing? And once you find it, once you’ve pinpointed the career of your dreams, the biggest hurdle of all is knowing the skills and training that will help you get there.

If you are dreaming of a new direction in their careers, or a new one entirely, Google is here to help with Career Dreamer.

Google gives you an AI powered career solution, where you can go and determine the skills and things you need to learn for your next dream role – all personalized to you.

The first step is going through the questionnaire, and then creating a custom prompt for you to use in Gemini to act as your career coach. (Copywriter note: Just don’t let it coach you into copywriting.)

It will even point you to the training sources you need, like Google Cloud Skills Boost and Google Career Certificates. Betcha can’t wait to put those on your LinkedIn profile!

Interested in learning more? Sure you are. Get Google Cloud certified in 2025—and see why the latest research says it matters.

32:27 Ryan – “This is way better than my usual method, which is complaining about something until they just give you that responsibility to make it your job to fix it, which is how I’ve advanced through my career.”

34:52 Enhancing AlloyDB vector search with inline filtering and enterprise observability

Google is introducing a new enhancement to help you get even more out of vector search in AlloyDB.

First, we are launching inline filtering, a major performance enhancement to filter vector search in AlloyDB.

Being able to perform vector search directly in the database, instead of post-processing on the application side, inline filtering helps ensure that searches are fast, accurate and efficient, automatically combining the best of vector indexes and traditional indexes on metadata columns to achieve better query performance.

Second, we are launching enterprise grade observability and management tooling for vector indexes to help ensure stable performance and the highest quality search results.

This includes a new recall evaluator, or built in tooling for evaluating recall, a key metric for vector search quality. You no longer have to build your own measurement pipeline and process for your apps to deliver good results.

38:30 Announcing Terraform providers for Oracle Database@Google Cloud

Google is sharing the GA of Terraform Providers for Oracle Database@Google CLoud. You can now deploy and manage Oracle Autonomous Database and Oracle Exadata Database Service resources using the Google Terraform provider. The release compliments the existing gcloud and google cloud console capabilities.

38:44 Justin – “I’ve always dreamed of being able to bankrupt a company with Terraform apply for my Oracle Exadata use cases. So thank you for that, Google. I really appreciate it.”

Azure

41:10 Announcing new models, customization tools, and enterprise agent upgrades in Azure AI Foundry

Azure AI Foundry is getting support for Open AI’s GPT 4.5 in preview on Azure Open AI. The research preview demonstrates improvements from scaling pre and post-training a step forward in unsupervised learning techniques.

Natural integrations with broader knowledge, higher “EQ” can help to improve coding, writing and problem-solving tasks

Accuracy and hallucinations: with lower hallucination rates (37.1% vs 61.8%) and higher accuracy 62.5% vs 3.8% compared to GPT-4o

Stronger human alignment improves the ability to follow instructions, understand nuance and engage in natural languages.

The latest wave of AI models from Microsoft Phi continue to push boundaries of what’s possible with smaller and more efficient architectures:

Phi-4-multimodal unifies text, speech, and vision for context aware interactions. Retail kiosks can now diagnose product issues via camera and voice inputs, eliminating the need for complex manual descriptions.

Phi-4-mini packs impressive performance in just 3.8 billion parameters with a 128k context window. Outperforming larger models on math and coding, and increased inference speed by 30%

Empowering innovation: The next generation of the Phi family

Stability AI Models with advanced generating techniques:

Stable Diffusion 3.5 Large

Stable Image Ultra

Stable Image Core

Cohere enhanced retrieval capabilities with Cohere ReRank 3.5

GPT 4.0 family expansion with Audio and Real Time preview

Plus you get all new customization tools

Distillation workflows

Reinforcement fine-tuning

Fine Tuning for Mistral

As well as support for bringing your own Vnet for AI Agent interaction and Magma (Multi-Agent Goal Management architecture) via Foundry Labs.

43:06 Ryan – “I do like the idea of those mini packs because I think that that’s that I’m more interested in that side versus the GPT 4.5 model. Like, cause I think that, you know, can have these giant mega models with all the information in them. But I mean, maybe it’s just my usage of AI is pretty simplistic too, but you know, their example of, know, being able to sort of take a, you know, different sets of information where it’d be visual text and then come up with a, like a repair program. Like that is, you know, like that’s the use case I’m more interested in versus just giant things. So that’s kind of neat.”

44:20 Announcing Provisioned Deployment for Azure OpenAI Service Fine-tuning

After fine tuning your models to make your agents behave and speak the way you like, you’ve scaled up your RAG apps – and now customers want it to be snappier and more responsive.

Luckily, Azure OpenAI service is offering (in preview) provisioned deployments for fine-tuned models, giving your applications predictable performance and predictable cost.

Provision throughput allows you to purchase capacity in terms of performance needs instead of per token. With fine-tuned deployments, it replaces both the hosting fee and token based billing of standard and global standard with a throughput based capacity unit called PTUs.

If you’re already using Provisioned Throughput units with base models, they work identically in fine tuned models and are completely interchangeable.

The two models you can provision deployments for are gpt-4o and gpt-4o-mini in North Central US or Switzerland with more regions coming in the future.

Note: if you want another region, click here and hit the “submit a request” button to get it considered for GA.

45:40 Matthew – “Well, that’s the problem; when you deploy your new app with a new thing, you’re like, OK, do I do provision? Do I hit my limits? And in Azure, and definitely some of the smaller regions or other regions than the primary ones like North Central, East US to those ones. You can hit those limits pretty easily and all of sudden then you get token limits or other errors that occur. So it’s like, you know, do you provision it or pay upfront, or do you risk a new feature of your app having an issue? Do you want your CFO yelling at you, or your customer?”

48:25 Announcing the launch of Microsoft Fabric Quotas

Microsoft has launched Microsoft Fabric Quotas, a new feature designed to control resource governance for the acquisition of your Microsoft Fabric Capacities. Fabric Quotas aims to help customers ensure that Fabric resources are used efficiently and help manage the overall performance and reliability of the Azure platform while preventing misuse.

Microsoft Fabric is a comprehensive service that offers advanced analytics solutions through multiple workloads, all available in a Single SaaS capacity model. Fabric is available via three skus:

Fabric Free trial: a time-bound per-user trial providing a capacity with a given size to every trial user

Power BI Premium: office-sold offers available as 12 month subscriptions

Fabric capacities: Azure PAYG offers available in multiple SKUs

Fabric Quotas limit the number of capacity units a customer can provision across multiple capacities in a subscription. The quota is calculated based on the subscription plan type and Azure region.

53:31 Availability metric for Azure SQL DB is now generally available

Azure SQL Database the modern cloud-based relational databases service is announcing the GA of Availability metrics for Azure SQL DBA enabling you to monitor SLA-compliant availability.

This azure monitor metric is at a 1-minute frequency storing up to 93 days. Typically, the latency to display availability is less than three minutes. You can visualize the metric in Azure monitor and set up alerts too.

Availability is determined based on the database being operational for connections.

A minute is considered downtime or unavailable for a given database if all continuous attempts by the customer to establish a connection to the database within the minute fail.

53:59 Justin – “If my database is down because I can’t connect to it for a minute, all of my app has failed. So I don’t, I don’t know that I need you to tell me that your availability was a miss. Cause I think I know from other reasons personally, but, like some customer somewhere must’ve just been like Microsoft, you have to tell us how available your database is. You promised this SLA and you don’t give us a way to measure it. And that’s BS. And that’s why this feature exists. And that’s the only reason why this feature exists because no one needs this unless you are being super pedantic.”

57:18 Native Windows principals for Azure SQL Managed Instance are now generally available

Azure is announcing the GA of Native Windows Principals in Azure SQL managed Instances. This capability allows the migration of Azure SQL Managed instances and unblocks the migration of legacy applications tied to windows login.

This feature is crucial for the SQL Managed instance link. While the managed instance link facilitates near-real time data replication between SQL Server and Azure SQL Managed instances, the read only replica in the cloud restricts the creation of Microsoft Entra principals.

With this new feature you have 3 authentication modes for SQL managed instances:

Microsoft Entra (default) this mode allows authenticating Entra users using Microsoft Entra user metadata.

Paired (SQL server default) the default mode for SQL Server Auth…. SA

Windows (New Mode): this mode allows authenticating Microsoft Entra users using the windows user metadata within sql managed instance.

59:02 Matthew – “I have feelings about this that I will not share because this podcast would never end.”

1:01:53 February 24th, 2025 Claude 3.7 Now Available in GitHub Copilot for Visual Studio

Last week we talked about Claude 3.7 shipping.

Well, it’s **good news**!

It’s available in Github Copilot now.

Closing

And that is the week in the cloud! Visit our website, the home of the Cloud Pod where you can join our newsletter, slack team, send feedback or ask questions at theCloud Pod.net or tweet at us with hashtag #theCloudPod

...more

View all episodes

By Justin Brodley, Jonathan Baker, Ryan Lucas and Matt Kohn