New website - feedback and ideas

[Jonathan Kay]

Email notifications come from:

noreply@invisioncloudcommunity.com

Could that be changed to something that identifies the Club?

Thanks

Jonathan

[anthonym]

On 31/12/2023 at 16:23, Andrew 21 Edney said:

(we don't want history rewritten)

we also don't want wrong/inaccurate/dangerous advice or to require reading of 100 page threads only to discover there is no solution - both of which is the current (crippled in my opinion) position. Adding more comments or requesting edits be made on our behalf are both impractical. 

A recent compromise I saw suggested, was to allow the opening poster to edit without limit - so I could edit very old posts I had been slowly fixing after being destroyed (ALL the links in that post are broken and i can no longer fix them) by the previous improved web site.

Edited August 7 by anthonym
add link to example

[anthonym]

It has become apparent to me that a.i. needs to be considered - it is an elephant in the room and ignoring it will not make it go away.

[Jonathan Kay]

1 minute ago, anthonym said:

It has become apparent to me that a.i. needs to be considered - it is an elephant in the room and ignoring it will not make it go away.

Discussing AI in ChitChat is fine. Posting the output of LLMs in TechTalk is a very bad idea. The provenance isn't checkable, and there have been multiple instances where the content has been shown to be totally wrong: "hallucinations". It could be dangerous.

Jonathan

 

[anthonym]

8 minutes ago, Jonathan Kay said:

Discussing AI in ChitChat is fine. Posting the output of LLMs in TechTalk is a very bad idea. The provenance isn't checkable, and there have been multiple instances where the content has been shown to be totally wrong: "hallucinations". It could be dangerous.

 

Sure you have a personal view about a.i. and so do I and everyone else. It seems to have missed your notice that I have NOT posted the output of LLMs in Techtalk and in addition to that I commented about using it which again you have ignored.

I suggest to you Jonathan that discussion and conversation are better attitudes than your apparent hostility and desire to control what I and others do.

"Posting the output of LLMs" this remark in itself indicates an innocence about how they can be used - for good not evil.

Anthony

 

Edited August 7 by anthonym

[Miker7]

I'm wary of LLM. On subjects I don't understand/know they seem to offer great advice / summaries. On subjects I do understand they sometimes offer bad advice/summaries. I appreciate humans can do the same ........ but you tend to know when a human does have experience in a subject so can trust their answer more. Added to this when I try to get a LLM to expand on an answer I seem to get more of the same original answer. I therefore like them to get a summary (and don't use them enough myself) but distrust them to do something technical.

[anthonym]

Methinks s chitchat debate would be interesting 

[anthonym]

delete duplicate

Edited August 7 by anthonym

[anthonym]

there's a famous case (literally) in my world when someone used a.i. to generate their tax case argument including case law. It turned out the two quoted cases were total hallucinations by the a.i. - case was lost!

and "no" I haven't checked if the story is true, but it could be and therein lies it's point.

Edited August 7 by anthonym

[rj]

1 hour ago, Jonathan Kay said:

Discussing AI in ChitChat is fine. Posting the output of LLMs in TechTalk is a very bad idea. The provenance isn't checkable, and there have been multiple instances where the content has been shown to be totally wrong: "hallucinations". It could be dangerous.

Jonathan

 

I can't see the problem as long as it's clearly marked. Better even if whoever posts it reviews the answer from the AI. Sometimes AI can find solutions I haven't thought of. That said I am VERY suspicious about the answers, but I certainly don't find it worse than asking the same question to the poster on more or less every single thread in TechTalk or providing links that don't answer the question - or where the answer is hidden on page 176 of 203, which often kills the discussion rather than being informative, and mostly keeps myself from posting here.

Edited August 7 by rj

[anthonym]

Question about the edit cripple: I am writing a blog entry about my dellorto carbs. It is VERY time consuming and may well take me more than a week to even reach it's first incarnation. Am I going to find i am denied access to my ongoing work because the (arbitrary) edit block activates?

Am I wasting my time completely trying to contribute anything while also hoping for the odd comment to help me along?

Should I be using Fakebook instead? Surely not...

Edited Friday at 15:51 by anthonym

[anthonym]

Is there a way to edit the blog tags I have made a right mess of?

[Golf Juliet Tango]

Anthony, can't you write it on your laptop first in Word (or Apple equivalent) and paste the whole article once complete? 

[anthonym]

sorry I missed your comment Stephen.

As regards writing first in some offline app that is usually what I do (Apple notes usually), however (frustratingly for some)  I nearly always make mistakes (whether typos (always) or content), and need to correct them. In addition most of what I write is on the edge or beyond what I understand, which is why I am writing about it (I am the student not the teacher in all matters 7) and the edit cripple means when I find I have written something that is wrong to whatever degree I cannot correct it for posterity (those reading several years later - or even myself checking what I did "last time"). 

I do understand that there are those who never make mistakes discovered later (long after 7 days etc) and everything they write will stand the test of time forever; I am not one of these. 

In addition I would suggest we should all be encouraged to add the solution to our question in the opening post AND to mark that o.p. as "Solved" so that future readers have immediate access to solutions instead of losing the will to live reading 156 comments only to find there is no solution - or if so it was never contributed. This latter being the problem with quoting links to old posts whose subject mentions the problem. 

However, I remain perfectly happy we have the team we have and view my constant edit cripple complaint in that light - things could have been much worse.. 😉 

p.s. AND blogs are also subject to this edit cripple - that's just plain daft.

I understand this is all driven by fear. I disagree with that fear.

Edited 10 hours ago by anthonym