Tag Archives: ai

One of those Gmail features that sound great on paper

Posted on Sep 1, 2021 by Henning Michael Møller Just

If you get a confirmation email for an event, a restaurant or hotel, Google will likely add this to your Calendar automatically. This sounds great on paper and once I found out, I disabled the feature for my account. Which … Continue reading →

Posted in Development | Tagged , , , , , , , , | Leave a comment

AI enhanced images

Posted on Aug 21, 2020 by Henning Michael Møller Just

Topaz Lab’s Gigapixel AI is currently one the best image-upscaling tools for the commercial market and the results are usually impressive. Sometimes a bit too impressive, such as “Gigapixel AI Accidentally Added Ryan Gosling’s Face to This Photo“: This reminds … Continue reading →

Posted in Development | Tagged , , , , , , | Leave a comment

Deep Learning is shallow

Posted on Dec 13, 2019 by Henning Michael Møller Just

Deep Learning, as part of Machine Learning and Artificial Intelligence, is actually shallow compared to human learning. Yes, DL uses more layers but it is still just patterns without explanation and will not be able to handle novel and unique … Continue reading →

Posted in Development | Tagged , , , , | Leave a comment

AI snake oil

Posted on Dec 10, 2019 by Henning Michael Møller Just

These slides from “How to recognize AI snake oil” by Arvind Narayanan does a pretty good job of putting AI – really Machine Learning, ML – capabilities in perspective. There’s a great example in them of machine learning being given … Continue reading →

Posted in Development | Tagged , , , , , , , | Leave a comment

Manual work is a bug

Posted on Jul 21, 2018 by Henning Michael Møller Just

In IT manual work is a bug and the fix is Document the steps Create automation equivalents Create automation Self-service and autonomous systems I would argue that this is also true for all manual office work, anywhere. The AI Lawyer … Continue reading →

Posted in Development | Tagged , , | Leave a comment

Google not being evil

Posted on Jun 7, 2018 by Henning Michael Møller Just

While Google was created with the words “Don’t be evil” as its foundation – later replaced by the parent company Alphabet’s “Do the right thing” – we have long known that the company no longer adhered to this slogan. I … Continue reading →

Posted in Development | Tagged , , , , , , | Leave a comment

Perception

Posted on May 27, 2018 by Henning Michael Møller Just

It’s funny that I have to spend time and energy to identify stop signs in a CAPTCHA, in order to prove that I am not a computer, while at the same time computers are driving cars around on roads trying … Continue reading →

Posted in Development | Tagged , , , , | Leave a comment

Successful spoofing of any kind destroys trust

Posted on May 11, 2018 by Henning Michael Møller Just

Four points and a note following the demonstration of Google Duplex: (soundcloud 1 & 2, or twistedsifter) 1. Is it OK to deceive people? As Jason Kottke writes in “Should our machines sound human?” this powerful demonstration is unnerving and … Continue reading →

Posted in Development | Tagged , , , , , , | Leave a comment

AI

Posted on Jan 20, 2018 by Henning Michael Møller Just

Hver eneste gang der er nogen der taler om hvor farlig kunstig intelligens (AI) er i sig selv, tænker jeg på nedenstående – jeg ville ønske jeg havde kunnet formulere det så kort og tydeligt selv: Let me say: too many … Continue reading →

Posted in Development | Tagged , , | Leave a comment

Computere og mennesker

Posted on Apr 9, 2017 by Henning Michael Møller Just

Efter at have set Seth Godin’s “23 things artificially intelligent computers can do better/faster/cheaper than you can” (vel vidende at han nok mener noget andet end AI) vil jeg godt citere den stadigt kontroversielle Heinlein: A human being should be … Continue reading →

Posted in Development | Tagged , , | Leave a comment