How to correctly perform automatic updates for Hermes agent
As someone who has used Hermes Agent for several months now, one of the biggest struggles is just keeping it updated, keeping it working through updates, and adapting the plugins I maintain to the frequent codebase changes.
A guide to AI prompting from a blindness perspective
A common use case that comes up when dealing with AI is alt(ernative) text generation for images. This is a loaded topic, and people have even questioned whether AI generated alt text is any better than no alt text at all? As a completely blind computer geek, I’ll give my own opinion briefly here: assuming you’re not proofreading the AI alt text for accuracy, no alt text is better than AI generated alt text. Why? Because if alt text exists, a blind reader might not realize it’s been AI generated, and may trust it to be more accurate than it actually is. However, if no alt text exists, screen reader users have tools that we can use to generate it for ourselves. We know how trustworthy our own tools are, and have probably gone to a lot of time and effort changing prompts and settings to get them to be as accurate as possible for the current state of technology.
AI Roleplay is a Good Thing, Actually
Not a week goes by without some publication releasing a new article, video, or paper on the completely real and valid harms AI is doing to the mental health of a large segment of society. Sometimes it's a weird news or human interest fluff peace on someone getting married to Chat GPT. Sometimes it's a deep, hard-hitting look at AI psychosis. Sometimes it's just encouraging everyone to point and laugh at lonely people with mental struggles who have AI girlfriends, so the news publication can make money hand over fist from the clickbait while victimizing the people struggling.
An Accessible Theme for Oh My Posh
Microsoft's Powershell has been my favourite shell for several years now. Not only is it an extremely powerful shell on Windows, it also runs perfectly well on mac and Linux; it tends to be the first thing I set up when configuring a new server, in fact. It's deeply configurable, integrates well with DotNet assemblies, and is well documented and logical. If you haven't explored powershell yourself yet, you absolutely should.
The Agent Solution I Actually Settled On
If AI is going to become a regular part of my workflow, I have a number of requirements for how it needs to work. Every single blind person has been burned by services locking down API's, making inaccessible updates, or just getting rid of functionality we depend on. In order to avoid history repeating itself, I want a system that is:
YubiKey And Screen Readers
Before we get into it, we should start with the basics for those who are unaware. What is a Yubikey, and why do you want one?
Openclaw Is More Fun to Fiddle With Than To Use
If you've never heard of Openclaw (formerly Moltbot, formerly Clawdbot), it's an open source AI tool that can actually do things in the real world. And sometimes, the things it does are actually the things you want it to do! Unfortunately, "sometimes" is doing a lot of heavy lifting in that sentence. Is Openclaw worth the massive hype it's received from people like Andrej Karpathy? I'll go over my experiences in this article.
The State of Modern AI Text To Speech Systems for Screen Reader Users
If you're not a screen reader user yourself, you might be surprised to learn that the text to speech technology used by most blind people hasn't changed in the last 30 years. While text to speech has taken the sighted world by storm, in everything from personal assistants to GPS to telephone systems, the voices used by blind folks have remained mostly static. This is largely intentional. The needs of a blind text to speech user are vastly different than those of a sighted user. While sighted users prefer voices that are natural, conversational, and as human-like as possible, blind users tend to prefer voices that are fast, clear, predictable, and efficient. This results in a preference among blind users for voices that sound somewhat robotic, but can be understood at high rates of speed, often upwards of 800 to 900 words per minute. The speaking rate of an average person hovers around 200 to 250 words per minute, for comparison.
Accessibility, the Origin of Innovation
Update: The Fable version of this material is now available here.
Correcting Years With NVDA and Espeak
I'm mostly documenting this for myself. I've been trying to switch away from eloquence for ages, with little success. One possibility is Espeak-ng, at least with NVDA on Windows, and on Android. Unfortunately, Espeak on IOS has a bunch of serious bugs, and hasn't been updated in forever.