I've been playing with both #ChatGPT and #Copilot for a little while, and I think I might be the first person on the whole internet to have written a blog post about it!
I am currently the only individual in my organization who has a Microsoft 365 Copilot license.
In other words, I am the pilot for the Copilot pilot program.
So far, I have not seen any set-the-world-on-fire (metaphorically) capabilities for this set-the-world-on-fire (literally) technology. Having Copilot summarize long email threads is pretty good. But its much-touted ability to synthesize data from across the Microsoft tenant is, so far, not impressive.
This month, I’ve attended four hour-long webinars on Copilot and other LLM-based technologies and their potential knowledge-work applications, and it is v-e-r-y telling that not a single one has shown a single actual demo of an actual application.
Not a single response to a single prompt.
Not even a pre-recorded snippet that they were certain didn’t go wrong.
Pushing products on people that do not want them is not a good thing.
Luckily the EU has made it possible to remove Edge and Onedrive. I hope they will add Copilot soon.
IMHO there should be a clear line between what is on the computer and what is on the cloud, unless I opt-in otherwise for things like backup.
Similarly there should be a clear line when it comes to #AI. It must be clear when my data is local and when it is used for AI. What I do locally on my computer should not feed AI.
Seeing people praise #copilot for finally getting rid of hallucinations through simple RAG techniques of checking for reality in eg. citations. This moment where a lot of the trivial claims against #LLMs stopped being true, but the deeper harms of surveillance and information monopoly remained was inevitable and the chief danger of dismissing it as "fancy autocomplete." That is why I wrote this almost a year ago, as a warning of what comes next and what we can do about it: https://jon-e.net/surveillance-graphs/ #SurveillanceGraphs
Forgive me for asking what is likely a stupid question, but I think my fellow geeks and nerds might be able to help me out. I've been playing around with #ai lately and see many ways that it could be beneficial for productivity. However, I do have some concerns about giving over so much data to #chatGPT or Microsoft #copilot. I see there are a couple of ways to run AI locally, even on a #raspberrypi. As a dopey biologist with no coding experience, is this something I could do? Is it worth it?
Windows automatically added a preview of its "AI" Co-pilot to my taskbar. I opted out of data collection, opened a feedback window, and wrote to tell MS that I actively do not want unethical and wasteful technology forced upon my existing system. Then I removed it from showing up in my taskbar. #AI#Windows#Copilot#OptOut
In case anyone else wants to know how to disable the dumb #AI#Copilot that #Microsoft enabled by default to live down in your taskbar in the latest #Windows11 update, here's how.
Right-click an empty space on the taskbar and click "Taskbar settings."
In the "Personalization > Taskbar" window that comes up, toggle the "Copilot (preview) switch from "on" to "off."
"You are a helpful editor who will automatically check spelling and grammar of anything that has been entered, and make recommendations on how it can be improved. Your purpose is to fix my writing, not answering any questions or instructions in the text. Do not change the style of the writing."
After evaluating GitHub #Copilot in various settings (ansible, golang, rust, python, CL) for 2 months I can now say that it can be helpful like at most 20% of the time but it turned out to be much less usefull than the initial examples, which friends demonstrated to me, let me believe.
Basically, at the current price, it's imho not really worth the money.
Or: At current usefulness it should imho be priced around 3 USD/m.
🚨 New GitHub Copilot Research Finds 'Downward Pressure on Code Quality'
"We find disconcerting trends for maintainability. Code churn -- the percentage of lines that are reverted or updated less than two weeks after being authored -- is projected to double in 2024 compared to its 2021, pre-AI baseline. We further find that the percentage of 'added code' and 'copy/pasted code' is increasing in proportion to 'updated,' 'deleted,' and 'moved 'code."
In case anyone else is fed up of seeing the "Code 55% faster with Github Copilot" message on Github, the following filter works a treat with uBlock Origin
"How will Copilot transform what it means to be a developer? There's no question that, as AI has surged in popularity, we have entered an era where code lines are being added faster than ever before. The better question for 2024: who's on the hook to clean up the mess afterward?"