?
On Sunday, my sister from Spain visited me again for a day before she had to return home. She was last in our old home town of Herzogenaurach, the "world city of sport"1Nikolas Pelke, New updraft for the Stub, 30 March 2023, Fränkischer Tag, https://www.fraenkischertag.de/lokales/hoechstadt-herzogenaurach/politik/diskussion-um-stub-nimmt-fahrt-auf-art-242803. adidas and PUMA are at home here, as well as many lesser-known and more well-known employers such as the automotive supplier INA Schaeffler. It is a tragedy that a German flagship city of this size such as Herzogenaurach, with its almost 14,000 daily commuters2 Andreas Scheuerer, Trotz Klimaschutz - Mehr und mehr Pendler im Landkreis Erlangen-Höchstadt, Fränkischer Tag, https://www.fraenkischertag.de/lokales/hoechstadt-herzogenaurach/verkehr/mehr-und-mehr-pendler-in-erlangen-hoechstadt-art-309755 has no public transport apart from a few ridiculous minibuses 🤯
I travelled from there by train as a child. Herzogenaurach had a railway station. But it was plastered over with tar to build car parks. A prime example of the failure of German environmental policy. I could puke. As is so often the case with me, I digress, another topic. Anyway, my sister was there as a guest of our parents. And then, on Sunday, for a day with me.
We had some very exciting conversations that day, where one thing led to another and then another. I found this so interesting that I want to try out a new format for my blog.
Let me expand a little to explain my idea. In and of itself, my eye control, which I rely on to communicate due to illness, is also a very limiting factor when writing my articles. The speed at which I can type with my eyes and the error rate leave a lot to be desired. What's more, my eyes tire quickly, dry out, turn red and are prone to inflammation3Patrick Ruppelt, Conjunctivitis or what, Where senses and mind meet, https://paddys.de/konjunktivitis-oder-was/. However, I have discovered one advantage of communicating in this way. If you first have to type each sentence on the tablet computer before you can click on the "Speak" button to pronounce the sentence, then the path to saving this sentence is not far away. At best, if I don't make a mistake, it's only two clicks away. It's easy to reconstruct the original conversation from all the snippets saved in this way. The only thing missing are the questions, but these are necessarily derived from the context.
To be honest, the thing that takes up the most time for me is quoting properly. My eye control is simply not made for that and no AI can really help me with that, because my database is my head. The limiting factor in teaching the AI what to do for me is, you guessed it, my eyes. I'll write about this in more detail and then link it here as soon as the article is finished. After all, the issue of missing source references has often been a "hot topic" in the past.
What AI can help me with, however, is to fulfil my own exaggerated demands for correct spelling and pedantically clean comma placement. When I write an article, I pay so much attention to it that I don't even proofread most texts. When it comes to the spoken word, in my case the "Geschreibsprochenen" (sorry grammar, that just sounds cooler than the grammatically more correct "geschrieben Gesprochene"), it's a completely different story.
My conversation partners spend more time waiting for my answers than actually talking to me. I try to compensate for this at least a little by not using capitalisation when speaking and replacing commas and full stops with double line breaks - because it sounds identical when read aloud by the software, but can be typed much faster using the eye control. Pronunciation, a chapter in itself. It seems to me that the motto in medical technology is "That'll do." in the sense of "What's good enough is just bad enough for us.". I can't remember ever hearing such bad pronunciation since Stephen Hawking's voice computer. May he rest in peace, a genius in his field.
When I talk to my mum in English, I write the text in Microsoft Word and use the "Read aloud" function. The English pronunciation of my voice computer is no longer just wrong like the German one is, but you simply don't understand a single word. In German, you have to have a very good command of the German language and listen carefully to understand everything.
Most of my carers would be completely at a loss if they couldn't read what I've written on the screen. Now I'm wondering what the point of a speech computer for the severely disabled is if the speech output is so incomprehensible that carers, doctors and even my lovely housekeeper can't understand anything. To be clear, my housekeeper even understands Bavarian. My speech computer, difficult. In everyday life, I don't use the voice output at all. Everyone is looking at my monitor anyway. A circumstance that can't be avoided for the reasons mentioned and which I hate like the plague and corona combined. For me, it's like people are looking into my brain and reading my thoughts.
It is sad that the voice output of Microsoft's "Word" writing programme, which has been included as a "minor matter" since version 20194Listen to your Word documents, https://support.microsoft.com/de-de/office/lauschen-auf-ihre-word-dokumente-5a2de7f3-1ef4-4795-b24e-64fc2731b001sounds infinitely better in all languages than that of specialised software that only has a single function - namely precisely this speech output. It was developed solely for this purpose, fulfils no other purpose and not even this one.
And so it is that when I write spoken texts, I deliberately misspell many words and sometimes entire sentences so that my stupid speech computer pronounces them approximately correctly or at least in a way that makes sense. So Bepanthen ointment becomes Beh pann tehn and when I ask Mummy for a coffee, something like Eid Fän ßieh ä Kapp off Koff ieh appears on the screen. Crazy, isn't it?
As nice as it is to save this text to the clipboard with two clicks to use it later for my blog, it is a pain to reuse this saved text later in my blog. Another thing that has not evolved in Word since the last millennium is the spelling correction. As soon as the first letter of a word or more than a single letter is wrong, the system fails. Crazy, isn't it?
This is where the AI finally comes into play. Unlike me, it only takes a few milliseconds for it to undo all my "deliberate mistakes" and give me back an error-free text in which I only have to insert my sister's questions. And the AI is pretty good at that itself. Crazy, isn't it?
In the next post, you'll find out why exactly this is one of the great strengths of "artificial intelligence" (AI) and what a dialogue reconstructed from snippets of conversation might look like. I don't want to keep you in suspense, it's just not quite finished yet and the topics that Sis has asked me about are far too exciting and extensive for me to squeeze them in at the end of this post. Anticipation is the greatest joy. It's going to be really good 😉
Episodes
