Alexander R Pruss
In typical lab research ethics, the following types of ethical considerations count against an experiment: risks of harm to the world outside the lab risks of harm to persons in the lab risks of pain to conscious entities in the lab. (The categories overlap: the persons in the lab are conscious...
Alexander R Pruss
In order to give valid consent, the subjects in an experiment need to be informed about all the harms that the experimenter plans to impose on them. Being deceived is a harm. Therefore, in order to give valid consent, the subjects in an experiment where the experimenter plans to use deceit need to...
Andreas Varotsis
NoteAbout this post This is an experiment on collaborating with AI: I wrote most of this blog post by collaborating with my Claw/Hermes Agent, running on a Raspberry Pi. I gave it a topic, we iterated on the draft together, and now it’s being published (mostly) without me ever touching the...
Andreas Varotsis
So, I really don’t like the scientific publishing ecosystem. I used to find the idea deeply inspiring and faintly aspirational, and then I had to actually interact with it, and came away with some confused blend of angry and profoundly disappointed. I think we can do better, and the good news is...
Fritz Meissner
Software delivery plans that consist purely of a list of tasks assume that learning is done before any code is written. This belief is false and leads to software project failure. Let’s imagine a plan that consists of five milestones. The organisation that wants this work knows far more about the...
danielmewes
Assume we want to use a vision-language model (VLM) to look at a given image and determine certain properties about it. Let’s say that, for the sake of the argument, we would like it to determine whether the image contains anything funny or not. As we all know, jokes are the most funny when you...
danielmewes
In my previous post about ThoughtNet, an attention-based neural architecture for variable-compute inference, I highlighted two limitations that I encountered with it: Slow and inconsistent convergence during training time Poor generalization on multiplication tasks, despite great...
danielmewes
The majority of today’s artificial neural network (ANN) architectures perform a constant amount of computation at inference time regardless of their inputs. This includes all recent GPT-style LLMs1 and other transformer-based architectures. Whether you ask an LLM to complete the series “1, 2,...
Ashley Newman
"the most productive part of the 30 minutes is the 30 minutes after" As I am writing this, I am completing the last day of my 30 minutes a day for 7 days challenge....
Ashley Newman
The other day I was visiting my parents in Hawaii when I complained to my mother about not wanting to do a planned run that afternoon. I told her that I was tired and was dreading physical...
Michael Allar
Today I was looking at the customer comments for my marketplace assets on the Unreal Engine Marketplace and found that across the entire site, usernames were being reported as "Unknown". For those who know my antics, I thought this was a perfect excuse to mess with the marketplace's JavaScript...
danieltyrkiel
You have an idea. You get excited with the potential, you see how it works out. You take small steps towards it, and then somehow you lose the drive. The idea wasn’t really that great. It could never happen, it’s stupid. You have another idea… The cycle could be a bit different for you. Your […]...
Michael Allar
In the middle of setting up a automated build server and I ran into an issue when trying to build dedicated servers for blueprint only projects. I basically needed to be able to auto-generate the *Server.Target.cs file for the project if its missing. I wrote an over engineered and bloated Node.js...