Anyone who has done publishing or worked with fonts would be aware of this issue. A mu in many fonts isn't available, and so you normally use the "Symbol" font, and in that font, the mu corresponds to a lowercase m. So if the text is then rendered without that font available, all "mu"s turn into "m"s. I expect that this hasn't been dealt with in the AI yet, hence the weirdness.

just wait for the first AI doctor to prescribe 125mg of digitalin when it should be 125μg.

 

'Common sense' is highly overrated. Common sense gets people into trouble. People use the term 'common sense' to justify stupid opinions.

In general, common sense refers to shared, unquestioned assumptions that form the Weltanschauung of the society the person lives in. They are the things that are "just right" or "just obvious" but usually can't be backed up with fact and logic because either they've never had to face that challenge or actually can't be because they are axiomatic to the world view.

Appealing to common sense is the tool of a scoundrel because it reaches into a person's foundational belief and instrumentalizes it. While it can be used for good (usually defensively "a good person just doesn't do X", etc.) precisely the same technique can be used to leverage inherent biases to turn people against each other.

The truth is, ChatGPT, et al, actually rely almost entirely on common sense since the output is dependent on the training set and, at least in the case of these commercial examples, they training set contains almost all the errors and biases of "the crowd". I think this approach is going to lead to an stealth intellectual mob rule—amplifying prejudice and canonizing commonly made errors so that we will necessarily consider them "correct" no matter what proof is brought against them.

A very minor and flawed example of this comes from language. It is minor because in and of itself it doesn't matter to the world, and it is flawed because language lacks "truth" in the sense people want it to have. That is, language necessarily and appropriately evolves to make communication more efficient or even possible when it comes to "new" ideas and things.

But this same mechanism can lead to degenerate outcomes, judged to be so by a reduction in the expressive power of the language. This can be in the form of a reduction in specificity or loss of information that is otherwise carried by a word. Words both denote and connote and the eloquent speaker selects for both. When connotation is lost because alternative words become indistinguishable synonyms, the language loses power.

This is not to say that a prescriptive dictionary is a good idea. Rather, it is an observation that the inevitable improvement in language in some places is unavoidably accompanied by decline in others. So, the foregoing said, an electronics-appropriate example meant to illustrate not condemn is possible.

I admit that I cringe every time a "latching" or "locking" connector is described as "buckling". The latch itself is called a "buckle". This appears to come from a machine translation of Chinese but since most connectors are now coming from China, it becomes the normal description. Because people learn languages from the speech (and writing) around them, many native English speakers now use this terminology.

Now, this doesn't harm the world and it is actually the way language goes—for good or bad as mentioned above. But, what is does do is point out that a source of information doesn't have to be "right" or "correct" to become authoritative. Eventually, no one will know what a latching connector is, they will tell you with great confidence that you mean a "buckling connector" and they would be "right" at that point.

In this way, pervasive use of AI language models to produce commercial copy (and this is happening right now) will amplify and canonize the error that matter much more than latch vs. buckle. Combined with the instantaneous and pervasive distribution of ideas via the Internet, we really don't stand a chance in preventing this pernicious degradation.

I believe the Internet will be seen as the main accelerator of the decline of civilization when the remnant reemerges and the philosophers and historians of that day examine what they can of how things came to be. Dark? Perhaps, but I haven't worked out how we can avoid this, at least not yet.

It's made me even more determined to live for the present and find ways to bring joy to my life and those around me, so that's a good side effect anyway...

[edited for typos without change to the content]

 

Anyone who has done publishing or worked with fonts would be aware of this issue. A mu in many fonts isn't available, and so you normally use the "Symbol" font, and in that font, the mu corresponds to a lowercase m. So if the text is then rendered without that font available, all "mu"s turn into "m"s. I expect that this hasn't been dealt with in the AI yet, hence the weirdness.

Well, I think this is quite an old outlook. For quite some time there has been almost no place that Greek characters, diacritics, and ligatures could not be used. It is ignorance or laziness that keep then out of text.

Thanks to UTF-8 and pervasive availability of the glyphs for all of these we've been able to include then for decades. Previously, the typewriter drove "special" characters out of text. They were special because they didn't appear on the typewriter. Even the use of underline rather than italic text is a typewriter artifact.

So, it wasn't for a lack of glyphs representing these things but because of ignorance (literally, not a value judgment) or laziness (again not a value judgment, people just didn't see a value in bothering with it, even if they knew how) that they weren't there.

And yes I have been involved in both publishing and the extensive use of fonts.

 

AI will become “intelligent” when it realises that what it finds on the internet is not necessarily the truth.

 

Anyone who has done publishing or worked with fonts would be aware of this issue. A mu in many fonts isn't available, and so you normally use the "Symbol" font, and in that font, the mu corresponds to a lowercase m. So if the text is then rendered without that font available, all "mu"s turn into "m"s. I expect that this hasn't been dealt with in the AI yet, hence the weirdness.

After i informed it, it replied that it understood and corrected it. I thought at least that was cool.
Now is it really an "It" or a "He" or "She" or "HeShe" ha ha.

 

Well, I think this is quite an old outlook. For quite some time there has been almost no place that Greek characters, diacritics, and ligatures could not be used. It is ignorance or laziness that keep then out of text.

Thanks to UTF-8 and pervasive availability of the glyphs for all of these we've been able to include then for decades. Previously, the typewriter drove "special" characters out of text. They were special because they didn't appear on the typewriter. Even the use of underline rather than italic text is a typewriter artifact.

So, it wasn't for a lack of glyphs representing these things but because of ignorance (literally, not a value judgment) or laziness (again not a value judgment, people just didn't see a value in bothering with it, even if they knew how) that they weren't there.

And yes I have been involved in both publishing and the extensive use of fonts.

Hi,

Thanks for explaining that.
I did find it interesting that once i told it what was wrong, it acknowledged the error and corrected it.

I often use the lower case English "U" for "micro" as in 5.4ua (5.4 microamps). I think a lot of people do that, and simulators accept that. I use lower case "M" for "milli" and simulators accept that. One thing that always bothered me about simulators is they dont like to use upper case "M" for "mega" they want you to write out "Meg" or possibly "meg". I guess they wanted a more clear distinction between "m" and "Meg" rather than "m" and "M".

 

AI will become “intelligent” when it realises that what it finds on the internet is not necessarily the truth.

That's an interesting view too. One thing i noticed one of them doing is providing a complete listing of where it got its information from, the web sites and links to the data.
Not sure if it can actually read data sheets like for transistors and IC's and stuff like that, but i think some of them can.

 

Hi,

Thanks for explaining that.
I did find it interesting that once i told it what was wrong, it acknowledged the error and corrected it.

I often use the lower case English "U" for "micro" as in 5.4ua (5.4 microamps). I think a lot of people do that, and simulators accept that. I use lower case "M" for "milli" and simulators accept that. One thing that always bothered me about simulators is they dont like to use upper case "M" for "mega" they want you to write out "Meg" or possibly "meg". I guess they wanted a more clear distinction between "m" and "Meg" rather than "m" and "M".

SPICE is so old that it came from the days when computers only worked in upper case.

 

Hi,

Thanks for explaining that.
I did find it interesting that once i told it what was wrong, it acknowledged the error and corrected it.

I often use the lower case English "U" for "micro" as in 5.4ua (5.4 microamps). I think a lot of people do that, and simulators accept that. I use lower case "M" for "milli" and simulators accept that. One thing that always bothered me about simulators is they dont like to use upper case "M" for "mega" they want you to write out "Meg" or possibly "meg". I guess they wanted a more clear distinction between "m" and "Meg" rather than "m" and "M".

The convention of Latin u for μ is common: uA, uP (microprocessor), etc. But for Farads, the practical fact that no component would be labeled in “millifarads” or “Megafarads” meant the ambiguity of mF was only theoretical. Things also had different common names, like micromicrofarad for pF.

On the other hand, for Ohms where there were milli- and Mega- values in common talk, you had megohms but for milliohms the m could be used because there was no common talk about “micröohms”*.

It is a relatively recent phenomenon that we are so careful about how we represent units and their prefixes. The regularity of the SI system, and its prescriptive nature were a reaction to the scattershot usages. Sometimes there might have been great consistency and regularity of usage in one group of users or another—but without agreement as to which version should be used. Usually though, it was a melange of versions driven by momentary convenience and misplaced orthodoxy, so SI has been a great thing, even if focusing on people as units is a bit dodgy.

—
* So here, in micröohms we have a purposeful, unconventional use of the dieresis to deal with the double o in the combined parts. The typewriter would have us use an endash—as opposed to its longer sibling emdash—to avoid the accidental blend that should not be pronounced. Today, the dieresis is omitted from words like cooperate because most people know it’s not like a chicken cooperate, but it began to be omitted because of the typewriter (and replaced with the endash when it was still needed).

There are other words with similar fates but people often think, if they are not sure of a spelling, that compound words which result in blends should get an endash. I would contend they should get a dieresis, but then I am a bit odd and really like diacritics which I find to be a nice embellishment when not strictly needed and the “right” way when needed as in our instant case.

Note, by the way, that “megohm“ is not ”megaohm” because the ao pair lacks the mellifluence of the go and people do have a natural desire for things that roll off the tongue. In fact, we will convert things into versions whose orthography appears to be nothing more than a convention once a word is polished by the figurative rock tumbler of common usage. Of course, it is the smoothed out pronunciation that is the convention while the spelling has a provenance (but, there is nothing sacred about orthography and this, too, is normalized and simplified through common use).

 

SPICE is so old that it came from the days when computers only worked in upper case.

That's funny because it's true. We had 5-bit mechanic teletypes as the I/O devices on computers used to program comm networks.

 

SPICE is so old that it came from the days when computers only worked in upper case.

Yes, and I think to some extent it is possible to alternatively say “programmers only worked in upper case” to a great extent since the operating system and protocol limitations on character encoding came from decisions by programmers. Even when we no longer used 5-bit codes and had no problem rendering the full upper and lower case range it took time for programmers to care enough to make the systems be able to, and to use, them.

 

That's funny because it's true. We had 5-bit mechanic teletypes as the I/O devices on computers used to program comm networks.

“Old minds think alike” (see: my response, as well)

 

Programmers stuck to upper case because it was ambiguous as to whether the machine on which they were working would consider ”A” and “a” as different or equivalent.
I only realised when the GNU compiler flagged it as an error that I had written “Variable” when I meant “variable”.

 

The convention of Latin u for μ is common: uA, uP (microprocessor), etc. But for Farads, the practical fact that no component would be labeled in “millifarads” or “Megafarads” meant the ambiguity of mF was only theoretical. Things also had different common names, like micromicrofarad for pF.

but mF is ambiguous, especially if you have an interest in old circuits, as it meant microfarad, usually, but not always, abbreviated to MFD.

* So here, in micröohms we have a purposeful, unconventional use of the dieresis to deal with the double o in the combined parts.

confusing the Germans as to the pronunciation (but not as confused a German Mötörhead fans)
https://en.wikipedia.org/wiki/Metal_umlaut

speaking of micro-ohms, I have just used a 1500A 60mV shunt in a project, that’s 40μΩ.

 

It will be interesting if they solve any problems in science, they will not need to be motivated. As we will provide the motivation. The real problem is what happens when if they become self aware?

 

It will be interesting if they solve any problems in science, they will not need to be motivated. As we will provide the motivation. The real problem is what happens when if they become self aware?

They cannot become self ware, that's the crux of all this. Science is often an exercise in reductionism, reducing some thing to lesser constituents, the belief is that a human mind can be reduced to an assemblage of simpler constituents.

The belief that self awareness, consciousness can be subject to reductionism is just that, a belief. In What Computers Still Can't Do, Dreyfus points out that there's no evidence at all that what the human brain does can also be done using symbol manipulation (which is all a computer is).

Software is algorithmic, it is the strict adherence to rules, so for consciousness to be possible in a machine we'd need to distinguish between unconscious algorithms and conscious algorithms and I know of no such distinction, I see no way to have rules that embody consciousness and rules that do not.

How could we design rules that when followed behave as if it is not following said rules? A machine that does not do what we design it to do is a broken machine!

 

I believe most animals are self aware to some extent, maybe even some plants. Since we cannot identify this quality it is impossible to say when something is self aware or not. It could very easily happen by accident.

 

Computer don't need to be self aware, conscious or intelligent. They just need to deliver things the masses desire cheaply and efficiently.

https://www.ftc.gov/business-guidance/blog/2023/03/chatbots-deepfakes-voice-clones-ai-deception-sale
Chatbots, deepfakes, and voice clones: AI deception for sale

Generative AI and synthetic media are colloquial terms used to refer to chatbots developed from large language models and to technology that simulates human activity, such as software that creates deepfake videos and voice clones. Evidence already exists that fraudsters can use these tools to generate realistic but fake content quickly and cheaply, disseminating it to large groups or targeting certain communities or specific individuals. They can use chatbots to generate spear-phishing emails, fake websites, fake posts, fake profiles, and fake consumer reviews, or to help create malware, ransomware, and prompt injection attacks. They can use deepfakes and voice clones to facilitate imposter scams, extortion, and financial fraud. And that’s very much a non-exhaustive list.

 

https://magazine.alumni.cam.ac.uk/t...Lent term issue&dm_i=6DCH,RPKP,1XTUOB,3GKQZ,1

 

but mF is ambiguous, especially if you have an interest in old circuits, as it meant microfarad, usually, but not always, abbreviated to MFD.

I don't understand this, it seems to be an objection but I addressed it directly:

But for Farads, the practical fact that no component would be labeled in “millifarads” or “Megafarads” meant the ambiguity of mF was only theoretical.

Because of the context of practical values mF, mf, MF, and MFD would be read as "microfarad" and if milli- or mega- were intended they'd have to be written out.

What am I missing in your reply?

speaking of micro-ohms, I have just used a 1500A 60mV shunt in a project, that’s 40μΩ.

Did you buy a 40μΩ resistor for it? My recollection of older resistors was labelling with "ohms" or Ω, and aside from K,k spelled out prefixes as needed, e.g.: megohms. I just realized I don't recall "kilohm" but there must have been some instances. Do you know if there were?