Beep boop - this is a robot. A new show has been posted to TWiT…
What are your thoughts about today’s show? We’d love to hear from you!
Regarding AI and the quality of its results, it can be wildly inaccurate. I had an experience with Bing Chat just after it was released, I wanted the lyrics to Mattemännken, it is a song sung by children on St Martin’s Day, they celebrate that here, instead of Halloween.
It is a simple, Platt German song, but Bing Chat came up with:
Mattemännken goue man Goue man, goue man Mattemännken goue man Goue man, goue man Mattemännken goue man Goue man, goue man
Wat häst du in de Tasse? In de Tasse, in de Tasse Wat häst du in de Tasse? In de Tasse, in de Tasse Wat häst du in de Tasse? In de Tasse, in de Tasse
Kaffee un Kauken Kauken un Kaffee Kaffee un Kauken Kauken un Kaffee Kaffee un Kauken Kauken un Kaffee
Dat will wi nich hebben Nich hebben, nich haben Dat will wi nich haben Nich haben, nich haben Dat will wi nich haben Nich haben, nich haben
Wat willt du denn hebben? Denn haben, denn haben Wat willt du denn haben? Denn haben, denn haben Wat willt du denn haben? Denn haben, denn haben
Äppel un Birnen Birnen un Äppel Äppel un Birnen Birnen un Äppel Äppel un Birnen Birnen un Äppel
Dat kanns du wol kriegen Wol kriegen, wol kriegen Dat kanns du wol kriegen Wol kriegen, wol kriegen Dat kanns du wol kriegen Wol kriegen, wol kriegen
Nu giff mi dat Äppelken Dat Äppelken, dat Äppelken Nu giff mi dat Äppelken Dat Äppelken, dat Äppelken Nu giff mi dat Äppelken Dat Äppelken, dat Äppelken
Nu giff mi dat Birnchen ook Dat Birnchen ook, dat Birnchen ook Nu giff mi dat Birnchen ook Dat Birnchen ook, dat Birnchen ook Nu giff mi dat Birnchen ook Dat Birnchen ook, dat Birnchen ook
Nu dank wi di so schön dafür So schön dafür, so schön dafür Nu dank wi di so schön dafür So schön dafür, so schön dafür Nu dank wi di so schön dafür So schön dafür, so schön dafür
Un wünsch di allns wat goot is Wat goot is, wat goot is Un wünsch di allns wat goot is Wat goot is, wat goot is Un wünsch di allns wat goot is Wat goot is, wat goot is
Un wenn wi wedder kommen Wedder kommen, wedder kommen Un wenn wi wedder kommen Wedder kommen, wedder kommen Un wenn wi wedder kommen Wedder kommen, wedder kommen
Dann willt wi mehr hebben Mehr hebben, mehr hebben Dann willt wi mehr hebben Mehr haben, mehr haben Dann willt wi mehr haben Mehr haben, mehr haben
The correct lyrics are:
Mattemännken gohe Man,
ob du uns was geben kann,
Appel odder Bireren,
Nörte sind vögierwen,
aule Statd - nigge Stadt,
leewe Jungfrau giff us wat,
laut us nich tou lange staun,
wi mört noch’n Hüsken wiedergaun,
bet nau Briärmen,
do wült wi us’n bittken wiärmen,
bet nau de nigge Stadt,
do stiärk’t se us oll tohaup innen Sack
I managed to find the correct lyrics using Duck Duck Go and then trained Bing Chat with the correct lyrics, so hopefully it now shows the correct result.
Luckily I knew roughly how the song went, so I could instantly see that what Bing was suggesting was utter jibberish. But it left me not trusting it with any answers, which kind of defeats the point, when you have to double check every answer it gives you.
In tests, I found that it gave results ranging from jibberish through bad to downright dangerous about 40% of the time. One of the worst was asking it a question about how much cash you can carry in the EU. It was out by a factor of 10, in the wrong direction! A direction that would have gotten me the rubber glove treatment by the German equivalent of the IRS, if I had followed Bing’s advice.
When you talk about the AI hype cycle, I think we must define what is meant by AI. If we think about Ai in a broader context, so beyond ChatGPT and other generative AI tools, we’ve had AI for ages. OCR, for instance, is a form of AI. I remember a couple of apps from about 2012 and 2013 - one was called Zoomreader, and I think the other one was called LookTel Money Reader. Zoomreader could do OCR on printed documents, made by the people who make Zoomtext, and LookTel Money Reader was one where you took a picture of a bank note and it would tell you what it is, i.e. £10, £20, etc. An app called TapTapSee did object recognition.
I suppose you could argue that screen readers, and speech synthesizers contain AI because, for the most part, they do not have a bank of words in them.
I agree that we need a CLEAR definition. I believe that most uses of the term these days are for hype rather than for categorization of the machine assistance. I don’t know what a clear definition is, I’m sure a committee would need to be struck, but it must contain something like “machine acting as a replacement for human intelligence, comprehension, decision making and effort.” If that were included, very little of what is today called AI would qualify.
It is a lot like the cloud or, at least, when “the cloud” first happened it just felt like a buzzword. To me, that’s kind of what AI is
1 Like
I think for most people at this point it’s a case of “I’ll know AI when I see it.”
I personally prefer the definition that AI is when a machine generates its own code based on a corpus of data (whether generated by humans or by itself as in the case of Aplha Zero) as opposed to a machine that acts on explicit instructions from a human programmer.
AI is machine made, code is man made.
Almost ALL of what we do on computers uses code generated entirely by human hands. What’s new, and interesting, is code that is generated by “machine learning” whether a LLM or GaN etc.
I’ll stick with that definition until I’m convinced otherwise.