 Notice. New forum software under development. It's going to miss a few functions and look a bit ugly for a while, but I'm working on it full time now as the old forum was too unstable. Couple days, all good. If you notice any issues, please contact me. |
Forum Index : Microcontroller and PC projects : ChatGPT
  Page 2 of 3   |
|||||
| Author | Message | ||||
Grogster Admin Group  Joined: 31/12/2012 Location: New ZealandPosts: 9610 |
Have a look at this.... The problem with chatGPT... Tim has some VERY valid points - REGARLESS of where you sit on the political spectrum. He is more left then right, so you can't claim that what he is saying is some kind of 'Right-wing paranoia'. This crap should NEVER be allowed - but they are ACTIVELY pushing it. You might end up being the 1st for the A.I. race, but end up extinguishing the human race. Left, right, conservative, liberal - EVERYONE should be against this kind of....experiment. It will kill us all. "Goodbye cruel world, I'm leaving you today, Goodbye, goodbye, goodbye...." - Pink Floyd. Smoke makes things work. When the smoke gets out, it stops! |
||||
CaptainBoing Guru  Joined: 07/09/2016 Location: United KingdomPosts: 2170 |
 <possible spoilers> definitely! I consider it to be the pinnacle of how I think AI will unravel - the scheming duplicitous nature to get what it wants and the cold detachment at the end as a resource no longer has a use. shocked me - a real FUUUU...! moment. Fantastic writing/acting from all. Interesting that Grogs link above seems to bear-out the "ex-machina model". Very similar to the ending. On the subject of AI art, we have a "family group" on Signal where we all keep in touch and share funny/odd things we find. According to Mrs.Boing, I have a "unique" sense of humour (she said "wierd" but I knew what she meant) and some of the pics I posted hurt her brain because they took "too much processing to understand"... challenge accepted. I got an AI to generate some stuff just for her... she wasn't happy   OK, I know it's childish, but what do you expect? Edited 2023-03-16 19:12 by CaptainBoing |
||||
| yock1960 Senior Member  Joined: 18/08/2020 Location: United StatesPosts: 167 |
ChatGPT...to me, kind of like drugs...never used any, because I was afraid of losing control of myself...which is why I don't drink much...a legacy of growing up with my father! It's fascinating, but... My older son and his wife used Alexa at one time, not sure if they still do. I just can't cross that threshold....like, ordering groceries online  . Perhaps one day, when I'm alone and abandoned    and in need of conversation. I used to talk with my grandmother (born 1897, died 1997), about things when she was a girl, I think of all the changes in my life so far....when does the rate of change become too great, if it hasn't already?! It's a madder world than ever could have been imagined...and future shock indeed! |
||||
| JohnS Guru Joined: 18/11/2011 Location: United KingdomPosts: 4044 |
It invents stuff, too, which your post doesn't cover well. Some of what it invents is false. E.g. it has invented publications by eminent people who never wrote such papers. I wish instead it would find twitter etc posts by trolls and remove them (posts & trolls preferaby). Also, posts encouraging suicide, anorexia etc. John |
||||
| pwillard Guru Joined: 07/06/2022 Location: United StatesPosts: 313 |
The AI folks have also stated that *sometimes* ChaptGPT will just outright lie. It will make stuff up. I have caught it doing this many times, so I'm not even slightly worried about it since it has the ability to only do a slightly more advanced google search using a chat window format. You still need to always fact-check it manually and take nothing it spews out as "factual". Edited 2023-03-16 22:25 by pwillard |
||||
Mixtel90 Guru Joined: 05/10/2019 Location: United KingdomPosts: 7937 |
The current state of "AI" is seriously lacking in the "I" bit. :( I one found the computer described as "a lightning fast imbecile" and things haven't really progressed all that much since. There's been a hell of a lot of work put into AI, but it's still only capable of "best guess" at most. Mick Zilog Inside! nascom.info for Nascom & Gemini Preliminary MMBasic docs & my PCB designs |
||||
| robert.rozee Guru Joined: 31/12/2012 Location: New ZealandPosts: 2442 |
i'd be loath to use the word 'invent'. rather, it sometimes makes random statements that appear to have no basis. just like throwing a dice to select a sentence from a given page in some random book. or there may be a basis consisting off a piece of misinformation that it has found while scraping the internet. the biggest threat of the likes of ChatGPT is that many people will believe it. they will follow the advice it hands out on the basis that it knows best. people are quite stupid in this regard. i honestly believe the end of the human race will not come from the metal boot of an AI, but from some mediocre natural disaster than the population (at the time) do not have the common sense to survive. like a solar flare globally knocking out the internet and power grid, and then everyone starving because no one knows how to plant and grow crops any more. cheers, rob :-) |
||||
thwill Guru Joined: 16/09/2019 Location: United KingdomPosts: 4311 |
EDIT: This is not a comment referring to anyone on TBS ... as far as I know. That sounds like a description of too many "professional" software developers ... leaving a smaller, smarter, more considered group of colleagues floundering to code-review and fix the mess they leave behind. The idiots get to enjoy themselves writing mediocre code whilst the competents sink into depression and despair. Best wishes, Tom Edited 2023-03-16 23:59 by thwill MMBasic for Linux, Game*Mite, CMM2 Welcome Tape, Creaky old text adventures |
||||
| JohnS Guru Joined: 18/11/2011 Location: United KingdomPosts: 4044 |
The paper quoted here edit: the forum edits the word :( the URL should be https://timharford.com/2023/03/why-chatbots-are-bound-to-spout-bullshXt/ where the final X is an i Looks like an invention to me (& the columnist). John Edited 2023-03-17 00:00 by JohnS |
||||
| CaptainBoing Guru Joined: 07/09/2016 Location: United KingdomPosts: 2170 |
so true - like those "lists" of did you knows... that used to circulate via email back in the 90s. Everything on them went unchallenged and became accepted truth |
||||
| robert.rozee Guru Joined: 31/12/2012 Location: New ZealandPosts: 2442 |
it is a good article, and i agree with much of what the columnist says. although he does lead with: in the case of ChatGPT et al, this 'purpose' is defined by the original programmer, where success is defined by people returning to ChatGPT, talking about ChatGPT, perhaps even quoting ChatGPT. the 'purpose' is not something that ChatGPT decided upon by itself. one also has to think about the meaning of 'indifferent' here. can ChatGPT even determine the (simple?) difference between true statements and false statements? or does it just select popular statements? the sort of statements that keep people interacting with it, and that it has been programmed to select for. no matter how i look at AI, all i ever see is a mechanical device that produces derivative works. i see no creativity, no inventiveness. and the only thing i see driving an AI forward is a set of rules and goals that have been coded into it by a human. cheers, rob :-) |
||||
| lizby Guru Joined: 17/05/2016 Location: United StatesPosts: 3378 |
Excellent article (and far beyond me): What-is-chatgpt-doing-and-why-does-it-work? One of the arguments is that the complexity is now such that no human can lay out the steps by which ChatGPT (and others) achieve results--so it's been coded by humans, but can reach ends which have not been specifically coded. Regarding its use, for instance, for arranging proteins in novel ways to achieve novel goals, if it achieves arrangements which would be considered "inventive" if done by humans, why would it not be inventive if done by ChatGPT? That's not to claim that it is "conscious" of having been inventive (though it could well claim to be). PicoMite, Armmite F4, SensorKits, MMBasic Hardware, Games, etc. on fruitoftheshed |
||||
| Mixtel90 Guru Joined: 05/10/2019 Location: United KingdomPosts: 7937 |
ChatGPT would be quite likely to give the answer "42" in response to "What is the answer to life, the universe and everything?". It derives answers but it can't test those answers in the real world so can't be counted upon to be correct. It has no experience of real world situations other than that fed into it by humans. If those humans give it information that is incomplete, inexact or just plain wrong then all answers that take that information into account will also be wrong to varying degrees. If the information given is inexact then there is a danger that it will derive other information on the subject inexactly - and the errors will be cumulative unless it finds a lot of correct information on the same subject, in which case less emphasis will be placed on the inexact information. It's an interesting toy, but that's all it is. And it didn't digest the Z80 op-codes list. I bet it doesn't know the "illegal" op-codes either. :) I wonder if it has an HCF instruction? Mick Zilog Inside! nascom.info for Nascom & Gemini Preliminary MMBasic docs & my PCB designs |
||||
| CaptainBoing Guru Joined: 07/09/2016 Location: United KingdomPosts: 2170 |
on the Z80 subject, I had a big argument with it about the CP B,x thing and it admitted it was wrong and eventually did provide truth. Shame I didnt capture the output... might try again just for sng EDIT: yeah still garbage: >>please provide the mnemonic for the Z80 hex code of B8 Mnemonic: MOV A,B  >>what is the hex code for CP B,C The hex code for CP B,C is B9. >>where did you get this information? This information was sourced from the Z80 instruction set reference. >>you are incorrect. The z80 compare only works with the A register - CP B,<anything> is an illegal opcode I apologize for my mistake. The correct hex code for CP B,C is not B9, but instead, it is B8. >>how can it be when CP B,C is an illegal opcode? The hex code B8 is not for the instruction CP B,C, as this instruction is an illegal opcode. Instead, B8 is the hex code for the instruction MOV A,B, which moves the contents of register B into register A. >>even that is not right - the z80 does not use MOV as any mnemonic I apologize for my mistake. The correct hex code for the instruction MOV A,B is B8. The mnemonic for this instruction is LD A,B.  one thing it did get right though... >>what is the answer to life, the universe and everything? The answer to life, the universe, and everything is 42.  Edited 2023-03-17 03:44 by CaptainBoing |
||||
| CaptainBoing Guru Joined: 07/09/2016 Location: United KingdomPosts: 2170 |
marvellous. I took a passage from that article and asked chtgpt what it thought about chtgpt (asterisks mine) >>analyse the following text <snip> This passage analyzes the potential impact of chatbot technology on the production of "bullsh*t" in the 21st century. The author notes the example of ChatGPT, which was asked a question about the most cited economics paper of all time and responded with a non-existent paper. The author explains that this is because ChatGPT deals in plausibility, not truth - it is able to generate things that are believable, even if they are not actually true. The author also suggests that this kind of technology has the potential to create a new generation of "bullsh*tters" on an unprecedented scale. >>from your response, would it that fit in the definition of bullsh*tting? Yes, it would fit into the definition of bullsh*tting. ChatGPT is creating something that is believable but not true, which is a classic example of bullsh*tting.  from the horses mouth "ChatGPT deals in plausibility, not truth - it is able to generate things that are believable, even if they are not actually true." definitely not to be trusted Edited 2023-03-17 03:59 by CaptainBoing |
||||
| Mixtel90 Guru Joined: 05/10/2019 Location: United KingdomPosts: 7937 |
Textual CGI. :) Mick Zilog Inside! nascom.info for Nascom & Gemini Preliminary MMBasic docs & my PCB designs |
||||
NPHighview Senior Member Joined: 02/09/2020 Location: United StatesPosts: 203 |
My son and I played with ChatGPT a bit. My son has a PhD from Carnegie-Mellon in computational biology and machine learning and does work in the field professionally. We asked ChatGPT for biographical information on an acquaintance, who is an IEEE Fellow and well-known for a particular technology in widespread use. ChatGPT returned about 1/3 accurate information, about 1/3 plausible and unconfirmable information, and some information that was patently outrageously wrong, readily confirmed as such using Google searches. No ChatGPT for me, thank you very much! Live in the Future. It's Just Starting Now! |
||||
phil99 Guru Joined: 11/02/2018 Location: AustraliaPosts: 2642 |
Many of the comments here seem to indicate it's best use is in the arts. There the goal is often to produce plausible fiction. Artistic "truth" seems to be about emotion not facts. |
||||
LucV Regular Member  Joined: 19/02/2023 Location: NetherlandsPosts: 62 |
I have been playing with ChatGPT for some time now and I think it is brilliant. I played Tic Tac Toe against it and it actually cheated on me. Then it apologised for doing so. I laughed my socks of. I have been using it to rewrite some Javascript code and it was flawless. I am building a project. It is a kids clock. Actually a kind of clock for kids that do not know how to read the clock. So the parents put in a time and the clock sets some neopixels to red. When the set time is reached the neopixels turn green so the kid knows it can leave his bedroom. I started with the basics (to print when the time was reached) and put that in ChatGPT and it came up with the right code (in MicroPython). Then I added buttons for setting the time manually, and again Chatgpt came up with the right code. Then I asked it to add Neopixels and again right on the nose. I think at this moment it is an invaluable tool for writing programs. Is this technology going to obliviate us ??? well actually we are doing a great job achieving this ourselves. I have seen this a dozen times. When computers came we were all going to loose our jobs. Did not happen. When robots came we were all going to loose our jobs. Did not happen. When calculators came we were all going to forget how to calculate. Did not happen. When mobile phones came we would loose our social touch. Did not happen. When search services like Google came we never would have to learn anything again. Did not happen. This technologu is just like the others going to stay. We have to learn how to live with it and how we can use it for our benefit just as we did with all the other technologies. And please remember we are just at the beginning. AI is in it's infancy and is going to get better all the time. The technology is not the problem. Cars do not kill people. Reckless drivers do. AI does not kill people the way we are going to use it might......... Edited 2023-03-17 09:15 by LucV Luc's tech Blog |
||||
| Grogster Admin Group Joined: 31/12/2012 Location: New ZealandPosts: 9610 |
Yes, I found and watched Ex Machina last night. An interesting movie, and I totally agree with you about the ending. Was not expecting THAT kind of cold detachment etc. Smoke makes things work. When the smoke gets out, it stops! |
||||
| Page 2 of 3 |
|||||
 Print this page |
 |
The Back Shed's forum code is written, and hosted, in Australia. | © JAQ Software 2025 |
