I don't know. I have no experience with it and probably lack the chops or inclination to use it, but my super-brain college-age nephew, who is into coding on the side, seemed to think it could spit out very individualised answers to coding questions, like "how would I go about coding this or that specfic thing". If an intelligent search engine was a hugely useful tool, then surely something that combines that with highly specialised, nearly complete seeds from which to start more complex tasks, almost serving like individualised tutorials (if not fully realised solutions), must be more than "what would your grandmother change on this tune? do that!" or whatever. I am now old and don't move in that space and am unable to evaluate the relevancy or (real, even positive) power of these things moving forward based on personal experience. I would honestly imagine most of us are. It's easy for me to focus on the considerable negatives as a result, and the meta/political side of it, where I feel more informed or wise based on past observation. I would like to hear more from people using it in concrete ways that go beyond spitting out expository text.penningtron wrote: ChatGPT is (currently) about as useful as oblique strategies: maybe a good way to stir up thought but then you spend almost equal time scrutinizing, expanding, and editing ideas.
Re: Thing: Artificial Intelligence
22Yeah, this is my field, and I haven't tried that yet either but it still lacks overall project context, and at the very least there's still a testing and review process (which in the near future might look like "you did more than just get ChatGPT to spit out this solution I hope?", really not much different than "you didn't just Google this did you.." is now). In that way I still see it as a fancy calculator and not THEY TOOK OUR JERBS but maybe I'm naive.jimmy spako wrote: Mon Mar 13, 2023 9:25 amI don't know. I have no experience with it and probably lack the chops or inclination to use it, but my super-brain college-age nephew, who is into coding on the side, seemed to think it could spit out very individualised answers to coding questions, like "how would I go about coding this or that specfic thing".penningtron wrote: ChatGPT is (currently) about as useful as oblique strategies: maybe a good way to stir up thought but then you spend almost equal time scrutinizing, expanding, and editing ideas.
(and yeah, this is just one small component of AI)
Re: Thing: Artificial Intelligence
23I asked the new Bing chat to compose a 3-voice fugue in the style of JS Bach, in E-flat major. This is what I got:
If it can help write code and cheat on an english lit assignment, whatever. I'll get excited when it can write a cantata that will win over the Prince Elector and secure me the position of Chat-AI Kapellmeister!
It could not produce the other two voice parts. So...whatever. One voice in a single octave without accidentals (no modulation?), just descending seconds. Fun!I apologize for the confusion. Here is the complete three voice fugue in the style of JS Bach in the key of E-flat major:
X:1
T:Three Voice Fugue in the Style of JS Bach
M:4/4
L:1/4
Q:1/4=120
K:Eb
V:1
z2 G2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A
If it can help write code and cheat on an english lit assignment, whatever. I'll get excited when it can write a cantata that will win over the Prince Elector and secure me the position of Chat-AI Kapellmeister!
Re: Thing: Artificial Intelligence
24It cannot listen to music. Makes sense it can't write it. I spent a decent amount of time asking it *specifically about music* by a certain artist. All it could give me was a short bio.Geiginni wrote: Wed Mar 15, 2023 7:36 pm I asked the new Bing chat to compose a 3-voice fugue in the style of JS Bach, in E-flat major. This is what I got:
It could not produce the other two voice parts. So...whatever. One voice in a single octave without accidentals (no modulation?), just descending seconds. Fun!I apologize for the confusion. Here is the complete three voice fugue in the style of JS Bach in the key of E-flat major:
X:1
T:Three Voice Fugue in the Style of JS Bach
M:4/4
L:1/4
Q:1/4=120
K:Eb
V:1
z2 G2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A
If it can help write code and cheat on an english lit assignment, whatever. I'll get excited when it can write a cantata that will win over the Prince Elector and secure me the position of Chat-AI Kapellmeister!
Re: Thing: Artificial Intelligence
25Great thread. The below is a copy of an email I send to our DVC last week. This is from the perspective of someone working in academia:
I am just so excited about ChatGPT. Not for reasons that most are, but still, I can barely contain myself. Honestly, I feel like I have been waiting for this moment. I know that this sounds a bit silly but I am being entirely sincere. Please bear with me and hear me out.
There is a threat to academic integrity posed by large language model-based generative AI such as ChatGPT. This much is obvious. If some students are tempted by and prepared to plagiarise content from the internet (and they are), or prepared to engage in contract cheating (and they are), then some students will be tempted by and engage in acts of academic misconduct enabled by ChatGPT and its cognates. Again, this much is obvious. What we might do to mitigate it is not.
There is a terrific opportunity here. It is wonderful, and it turns on relationships. Academic integrity is only one type of many. It sits next to personal integrity, and is an instance of professional integrity more generally. It will include pedagogical and research integrity as its members. Others might include emotional integrity, artistic integrity, social integrity, managerial integrity, and more.
I do not know the taxonomy of integrity in detail, but I do know that it is always founded upon a relationship of some sort - between a person and some other thing. Sometimes the other thing is another person or persons. Other times it is responsibilities, other times office, and so on. The point here being only that anything that threatens such relationships will threaten the integrity founded upon them.
Academic integrity is founded upon relationships between members of the academic community, and this includes our students of course. What might threaten these relationships? Rather a lot of things I think! Good relationships require vigilance after all. But what I have in mind here is standardisation. Standardisation threatens these relationships because standardisation is, by definition, impersonal. And now to close the circle finally - standardisation begets alienation, and alienation eats away at our interpersonal relationships.
Here then is the opportunity afforded to us by ChatGPT and its ilk as I see it. Our opportunity is to bring discussions exploring authenticity into the daylight. It is to recognise and articulate the dangers posed by the alienating service-model understanding of the University. It is to reflect on the real, relationship-building and integrity-preserving meaning of authenticity. It is to finally emerge from lazy post-modern semantic-nihilism, and Resurrect the Author.
The opportunity more broadly is to reflect upon our pedagogical practice, minding constantly that we are creating persons by virtue of what we ask them to do. We are what we do - along with our reasons for so doing - so what we ask of our students will create the persons into which they turn.
Now that ChatGPT lives among us, we have reason to bring to the fore and cultivate those aspects of our students and ourselves that cannot be automated or standardised, and I cannot wait!
I am just so excited about ChatGPT. Not for reasons that most are, but still, I can barely contain myself. Honestly, I feel like I have been waiting for this moment. I know that this sounds a bit silly but I am being entirely sincere. Please bear with me and hear me out.
There is a threat to academic integrity posed by large language model-based generative AI such as ChatGPT. This much is obvious. If some students are tempted by and prepared to plagiarise content from the internet (and they are), or prepared to engage in contract cheating (and they are), then some students will be tempted by and engage in acts of academic misconduct enabled by ChatGPT and its cognates. Again, this much is obvious. What we might do to mitigate it is not.
There is a terrific opportunity here. It is wonderful, and it turns on relationships. Academic integrity is only one type of many. It sits next to personal integrity, and is an instance of professional integrity more generally. It will include pedagogical and research integrity as its members. Others might include emotional integrity, artistic integrity, social integrity, managerial integrity, and more.
I do not know the taxonomy of integrity in detail, but I do know that it is always founded upon a relationship of some sort - between a person and some other thing. Sometimes the other thing is another person or persons. Other times it is responsibilities, other times office, and so on. The point here being only that anything that threatens such relationships will threaten the integrity founded upon them.
Academic integrity is founded upon relationships between members of the academic community, and this includes our students of course. What might threaten these relationships? Rather a lot of things I think! Good relationships require vigilance after all. But what I have in mind here is standardisation. Standardisation threatens these relationships because standardisation is, by definition, impersonal. And now to close the circle finally - standardisation begets alienation, and alienation eats away at our interpersonal relationships.
Here then is the opportunity afforded to us by ChatGPT and its ilk as I see it. Our opportunity is to bring discussions exploring authenticity into the daylight. It is to recognise and articulate the dangers posed by the alienating service-model understanding of the University. It is to reflect on the real, relationship-building and integrity-preserving meaning of authenticity. It is to finally emerge from lazy post-modern semantic-nihilism, and Resurrect the Author.
The opportunity more broadly is to reflect upon our pedagogical practice, minding constantly that we are creating persons by virtue of what we ask them to do. We are what we do - along with our reasons for so doing - so what we ask of our students will create the persons into which they turn.
Now that ChatGPT lives among us, we have reason to bring to the fore and cultivate those aspects of our students and ourselves that cannot be automated or standardised, and I cannot wait!
"lol, listen to op 'music' and you'll understand"....
https://sebastiansequoiah-grayson.bandcamp.com/
https://oblier.bandcamp.com/releases
https://youtube.com/user/sebbityseb
https://sebastiansequoiah-grayson.bandcamp.com/
https://oblier.bandcamp.com/releases
https://youtube.com/user/sebbityseb
Re: Thing: Artificial Intelligence
26If you read Pharmako-AI (particularly the bit about the poem), you'd find that email really funny. And that's a good thing.
at war with bellends
Re: Thing: Artificial Intelligence
27I haven’t directly used ChatGPT myself, but my colleague recently uploaded a draft sell sheet to it and asked it to check spelling and grammar. It did way better of a job than any other grammar/checker that I’ve seen.
Which makes me wonder if stuff that gets uploaded to ChatGPT gets cycled into its hive mind, resulting in it getting “smarter”. (Admittedly, I don’t know the finer details of how ChatGPT works, but this is an essential aspect of machine learning... you need to provide input to AI and check its output to continuously improve it.)
AI can and will be used for good and bad. Don’t blame the technology, blame the user. NC
Which makes me wonder if stuff that gets uploaded to ChatGPT gets cycled into its hive mind, resulting in it getting “smarter”. (Admittedly, I don’t know the finer details of how ChatGPT works, but this is an essential aspect of machine learning... you need to provide input to AI and check its output to continuously improve it.)
AI can and will be used for good and bad. Don’t blame the technology, blame the user. NC
jason (he/him/his) from volo (illinois)
Re: Thing: Artificial Intelligence
28Current consumer usage seems to be faddish. Like a photograph filter that mostly works, but still fucks up one eye in the pic or something.
Pending improvements before applying judgment.
Pending improvements before applying judgment.
Re: Thing: Artificial Intelligence
29Yeah, not too surprising. It's basically an overgrown spell-checker app. Not of much use or interest to me. Won't be taking over the world anytime soon. Maybe vexing to untenured faculty without more important things to worry about.enframed wrote: Thu Mar 16, 2023 12:24 amIt cannot listen to music. Makes sense it can't write it. I spent a decent amount of time asking it *specifically about music* by a certain artist. All it could give me was a short bio.Geiginni wrote: Wed Mar 15, 2023 7:36 pm I asked the new Bing chat to compose a 3-voice fugue in the style of JS Bach, in E-flat major. This is what I got:
It could not produce the other two voice parts. So...whatever. One voice in a single octave without accidentals (no modulation?), just descending seconds. Fun!I apologize for the confusion. Here is the complete three voice fugue in the style of JS Bach in the key of E-flat major:
X:1
T:Three Voice Fugue in the Style of JS Bach
M:4/4
L:1/4
Q:1/4=120
K:Eb
V:1
z2 G2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A2 G2 | F2 E2 | D2 C2 | B2 A2 | G2 F2 | E2 D2 | C2 B2 | A
If it can help write code and cheat on an english lit assignment, whatever. I'll get excited when it can write a cantata that will win over the Prince Elector and secure me the position of Chat-AI Kapellmeister!
Re: Thing: Artificial Intelligence
30I've noticed that if you have multiple conversations on similar topics that its response gets more tailored, so there's some kind of relational aspect there.jfv wrote: Thu Mar 16, 2023 6:45 am Which makes me wonder if stuff that gets uploaded to ChatGPT gets cycled into its hive mind, resulting in it getting “smarter”.
Plus Moore's law - where this thing will take us in a few years is beyond our imagination.