My team just launched a new "less BS" #AI page. Check the three links there with FREE GitHub (OSS MIT Licensed!) COURSEWARE on building apps, learning ML, and learning how to pair-program with Copilot https://developer.microsoft.com/ai
Comments
Log in with your Bluesky account to leave a comment
Really disappointing in your pushing this ai garbo.
I know your deep in Microsoft, but for a long time, with open source and promoting increased diversity you appeared to be a good one. Guess the evil gets everyone in the end.
Recommend checking out this fantastic session from @scott.hanselman.com at Scottish Summit in October - really great overview of responsible AI and a good AI explainer, Scott is defo one of the good guys.
Try exploring the talks that I’ve given on the topic first before you declare chicanery. I talk about open models and open weights, I show local models first and open source tooling, I talk about responsible AI and ethically sourced data, and focus on ways to remove bias. I’m not an AI bro *at all*
I am many things, but I’m not evil and I don’t need the money. AI can be super dangerous and if someone’s gonna be talking about it, I’d rather it be me or someone on my team rather than some grifter who is gonna shill “AI all the things” like it’s crypto
Having seen you speak on the subject, I can say it was one of the most thoughtful talks I’ve heard. Quite masterful in the way you were able to highlight the responsibility each dev has.
I’m absolutely permitted to express my opinion and if they wanna fire me for it they’re welcome to do so.
I talk about mitigating the environmental impact in all my talks and material and I I focus on using the absolute smallest model for the work and very much prefer bias towards tiny local models
I also don’t use video or image generation because the environmental impact is too great. Text generation and text summarization is the lowest footprint by a couple of orders of magnitude
The LinkedIn mass pivot from 'crypto used car salesman' to 'AI used car salesman' told me I needed to be very cautious about AI long before I did the reading that I ended up having to do when the shillery leaked in to my D&D group
Yeah, I mean for better or for worse it's here to stay. Better to understand the capabilities and constraints to make informed decisions in favor or against.
That’s my perspective in fact. I actually gave a full workshop to the educators at our local school district, with a focus on reducing and removing bias as well as the ethics of the corpus
I grant you that Evil was perhaps hyperbole! but anyone advocating the use of generative AI for anything is not on the side of good, or that interested in having a planet to live on. I only posted because Scott is so thoughtful about so many things maybe he is saveable.
The internet does have a lot of infrastructure, and uses a colossal amount of resources. but the use of that infrastructure is (broadly) for the good, and commensurate with the utility.
The training of LLM's takes such a massive amount of compute that coal and gas power stations are turned on for it. the utility is spam websites, voice matched scam calls to your gran. all from the stolen work of artists with no possibility of them being able to grant rights or get compensation.
Comments
I know your deep in Microsoft, but for a long time, with open source and promoting increased diversity you appeared to be a good one. Guess the evil gets everyone in the end.
https://youtu.be/nxAxsFSVHHM?si=vFuprDQcwykPQeG1
I appreciate that you may not be permitted to express your full opinion due to corporate associations.
I talk about mitigating the environmental impact in all my talks and material and I I focus on using the absolute smallest model for the work and very much prefer bias towards tiny local models
Mastering GitHub Copilot for Paired Programing
https://microsoft.github.io/generative-ai-for-beginners/#/03-using-generative-ai-responsibly/README?wt.mc_id=academic-105485-koreyst