James Croft

Commanding AI programmers, realizing your creativity - James Croft Explore how commanding AI programmers fosters creativity and efficiency for developers harnessing the power of your imagination.

The role of developers is evolving, whether we want to accept it or not.

Today, we can realize the potential of commanding AI programmers to solutionize. The only limit... your creativity!

www.jamescroft.co.uk/commanding-a...

A cat sleeping in the top drawer of a set of IKEA office drawers in with a comfy blanket bed inside.

Decided to convert the top drawer in my IKEA office drawers into a cat bed... Daisy instantly took a liking to it 😂

It's a tool, tools have a purpose. Use generative AI when gen AI is the best fit for the task. Use discriminative AI when discriminative AI is best for the task. Use deterministic logic when deterministic logic is the best for the task.

Some `if` statements just don't need to be replaced w/ models.

A card with a light green background describes the Theft Protection feature for Android devices, stating 'Use built-in safeguards to protect your personal data if your device is stolen.' An illustration shows a hand holding a phone. A toggle switch labeled 'Theft Detection Lock' is shown, with a description that reads, 'The screen will lock if this setting detects that someone took your device and ran away.'

Turns out my phone got stolen...

I had 15 minutes to get across Schiphol Airport to catch my connecting flight and I ran like the wind with it!

Love this new Theft Detection feature that locks your phone if someone steals it, then moves really fast away.

It worked, but not as intended 😂

The image displays the results for James Croft, who participated in the "Run for Wildlife Night 5K" on January 4, 2025. The event's logo is prominently featured at the top, showing a crescent moon, stars, and silhouettes of animals and runners. Below, the details for James Croft are as follows: Bib Number: 344 Gender: Male Age Category: Ages 30-34 Status: Finished Finish time: 00:31:48 Overall Ranking: 214 out of 675 Age Category Ranking: 11 out of 23 Gender Ranking: 114 out of 204

Started the year off with a night time winter light trail run!

Not a PB but thoroughly enjoyed the route 🏃🏅

I'd take this user experience every time

🚀 I created this new quickstart template providing you everything you need to setup a new Azure AI Content Understanding project in the Azure AI Foundry, complete with Microsoft Entra ID authentication.

Get the template here 👉 github.com/jamesmcroft/...

#Azure #AI #ContentUnderstanding

Introducing Phi-4: Microsoft’s Newest Small Language Model Specializing in Complex Reasoning | Microsoft Community Hub Today we are introducing Phi-4, our 14B parameter state-of-the-art small language model (SLM) that excels at complex reasoning in areas such as math, in...

Introducing Phi-4: Microsoft’s Newest Small Language Model Specializing in Complex Reasoning.

https://techcommunity.microsoft.com/blog/aiplatformblog/introducing-phi-4-microsoft%E2%80%99s-newest-small-language-model-specializing-in-comple/4357090
#ai #aimodels #phi4 #slm #azure #azureaifoundry

A screenshot of the Azure OpenAI Image Token Calculator, designed to calculate token usage and costs for image processing in Azure OpenAI services. It highlights the preselection of an A4 preset at 3508x2480 pixels highlighting the cost.

QOL improvement to the unofficial Azure OpenAI Image Token Calculator!

You can now select presets for common print and web image sizes. Want to work out how many tokens a 10-page document would cost to process with GPT-4o? Now you can!

jamesmcroft.github.io/openai-image...

Possibly Chicago too 😅

So who wants a verified 'Microsoft' and 'Microsoft MVP' label on their profile and all the posts?

I just finished setting up @bluesky.ms as a labelling service.

Go subscribe to the label to start seeing labels on verified MVPs and Microsofties.

🧵👇

👋👀

I asked the electrician:
- How bad is the previous installation?
He replied:
- This isn’t that bad, things change over time, things get old. You measure a good implementation on how easy it is to replace. This one is easy to replace.

I love that way of thinking, it relates to programming as well.

I paid to watch origami on a premium TV channel.

It was paper view

I work with colleagues and customers from all over the world whose first language isn't English. Teams captions is a life saver in helping me to have better conversations with them 😊

We are beyond excited to announce the Service Bus local emulator! This has been the most requested feature for Service Bus, and today is the day.

You can find the official announcement at techcommunity.microsoft.com/blog/messagi....

My team has had a lot of fun working on the AI Dev Gallery and we are excited to share it with the world. Experiment with models and local AI in your Windows apps - get it here: github.com/microsoft/ai...

#WindowsApps #AI #DevTools

But what are your thoughts on Coke Zero? 🤔

Screenshot of tweet. A software developer's confidence throughout their life: 0 years experience: I have no idea what's going on. 5 years experience: I AM A SENIOR DEVELOPER I CAN BUILD ANYTHING GET OUTTA MY WAY 10/15 years experience: I like my personal life. 20 years experience: I AM AN IMPOSTOR

This is probably my favorite tweet.

It was popular, but it was more than that. It described my mental state at the time - I was struggling a bit in my career. And I spent like a day workshopping it.

A tortoiseshell cat watching the Microsoft Ignite keynote live stream

I'm not the only one fascinated with the advancements in security for AI #MSIgnite

Microsoft Ignite 2024 Book of News The Book of News is your guide to all key items being announced at Microsoft Ignite 2024.

There are a ton of announcements across soooo many products and services. We put all the amazing #MSIgnite news in the Book of News.

news.microsoft.com/ignite-2024-...

A diagram showing how to generate insights with the Azure AI Content Understanding service. It shows audio, video, document, and image inputs being understood with content extraction, field extraction, and confidence scores, to produce structured insights and new content outputs.

Excited about the announcement of our new Azure AI Content Understanding service! A content processing pipeline that can take documents, audio, and video to generate structured, actionable insights with confidence to reduce human intervention!

techcommunity.microsoft.com/blog/azure-a...

#MSIgnite

Microsoft Ignite | November 18-22, 2024 | Chicago, IL Join us at Microsoft Ignite on November 18-22, 2024 in Chicago, IL and online to achieve more with the power of the Microsoft Cloud, Microsoft Copilot, and AI.

It's time for the Microsoft Ignite keynote! Join us online now 😄 ignite.microsoft.com/en-US/home#s... #MSIgnite

Azure Container Apps dynamic sessions general availability, and more | Microsoft Community Hub Python code interpreter and custom sessions are now GA. JavaScript code interpreter in public preview. Microsoft Copilot uses Azure Container Apps dynamic...

Azure Container Apps dynamic sessions updates at #MSIgnite

✨ Generally available: Python code interpreter
✨ Generally available: Custom container
✨ Public preview: JavaScript code interpreter

Plus: Dynamic sessions powers Microsoft Copilot

aka.ms/Ignite24/Blo...

Can you try again? It seems to be working. We recently had an update to Tech Community so you _may_ need to clear your cache

As a sneak peek, you can achieve highly accurate, confident, and cost-effective extractions using GPT-4o with its vision capabilities!

And small language models like Phi-3.5 MoE aren't far behind in terms of accuracy compared to the heavy hitters.

Evaluating the quality of AI document data extraction with small and large language models Evaluating the effectiveness of AI models in document data extraction. Comparing accuracy, speed, and cost-effectiveness between Small and Large Language...

If you're exploring GenAI for extracting structured data from unstructured or structured documents for downstream processing, I recommend checking out my updated article evaluating various techniques and models.

techcommunity.microsoft.com/blog/azurefo...

#Azure #OpenAI #Phi

They were very good from what I remember. Not sure I want to try squirrel to compare though