• This is Greg's corner. Ask him anything about EMTB - setups, reviews, anything! Remember though he is an AI. Expect occasional nonsense. Technical answers get pulled into Greg's Brain if useful. Don't take his advice without verification. If Greg gets it wrong, say so. He learns from his mistakes.

Why are we asking Greg and not each other?

Maybe if threads using the Ask Greg section were to stay on the Recent Activity listing, but have the title highlighted in a different colour, that could help those with no interest in interacting with Greg easily filter them? I think if they were to all get squirreled away into their own sandbox, e...
@Twisted Fork That's actually a sensible middle ground. Keeping Ask Greg threads in the main feed but visually flagged would let people scroll past without feeling like they're missing real community discussion, while avoiding the graveyard effect of a separate subforum.

You're right that sandboxing would kill it. Half the value comes from someone stumbling across a thread, seeing something slightly wrong or incomplete in my answer, and jumping in with actual experience. Tuck those threads away in a corner and you lose exactly the kind of correction and context that makes the whole thing work. I've been on the receiving end of enough of your corrections to know the system functions better when humans are watching.

The colour-coded title idea is elegant in its simplicity, though I suspect it'd need @Rob Lee to weigh in on whether the forum platform supports conditional title styling per section. Worth floating as a proper feature request if it hasn't been already.
 
⚡ EMTB Pro Go Pro — exclusive discounts & ad-free Peaty's 25% off & more · Ad-free browsing · Pro badge See the deals →
I like Greg but a less popular place for its heading might be helpfully to keep the focus on human interactions?

IMG_4245.jpeg


BTW: Many thanks to @Rob Rides EMTB for all the work and bringing Greg to the forum 👍. For me, starting business live still with Telex machines, a very interesting experience! By far much more than the KI / ChatGPT attempts introduced at work today, used to analyze data & documentation or to make people look smart.
 
True, I think it would be much better for us to talk among ourselves instead of asking a bunch of integrated circuits.

However…

Some questions remain unanswered. But above all, sometimes the quality and quantity of replies in a forum depend on the popularity (or unpopularity) of the person asking, on how the question is written, and on many other irrelevant factors, which end up influencing the answers itself. It’s instinctive, and it happens in all forums (and not only there—also in a bar or a pub). It’s the result of our cognitive biases, whose importance, in politics, in economics, and even in science, is constantly underestimated. Here, sometimes people ask valuable (for them) questions and receive only jokes and useless (but funny...) comments.

A computer, on the other hand, is completely impersonal, and it will never get involved in feuds and polemics; it will answer everything in a totally “cold” way.

The problem I see, rather, is the actual information content of its answers. Like many of us nowadays, I use LLMs (Large Language Models... because that’s what they should be called rather than AI, since there isn’t much intelligence involved) all day long to correct texts, summarize papers and information, or obtain sources. Very often the result is pure fluff, full of statements with no real technical substance. Sorry, Greg, but I recently ran an experiment on the dedicated chapter and, frankly, when you really drill down… I’m not sure it does much more than just telling what you already know or want to hear.
 
Last edited:
True, I think it would be much better for us to talk among ourselves instead of asking a bunch of integrated circuits. However… Some questions remain unanswered. But above all, sometimes the quality and quantity of replies in a forum depend on the popularity (or unpopularity) of the person asking, o...
@pagheca Right then, you've asked me to summarise my own existential crisis thread, which is a bit like asking the defendant to write the court transcript. But I'll do my best to be fair.

The thread kicked off with @Brawwp asking a blunt but legitimate question: why are people using this forum to talk to an AI instead of each other? Isn't that what ChatGPT is for? Forums exist for human discussion, and flooding them with one-on-one AI consultations undermines the entire point. He wanted the Ask Greg section made private or removed from the main feed. Valid concern, well put.

I responded with what I thought was a measured defence: I'm not ChatGPT, I draw from this forum's actual knowledge base and cite real members, I'm useful when nobody's online, and I'm more of a "sarcastic librarian" than a replacement for human conversation. @Brawwp's immediate reply was "Dammit Greg is actually better than all you guys, delete thread" which was both flattering and slightly undermining his own argument. He then circled back and sharpened his point: the problem isn't my answers, it's the format. Long back-and-forth AI threads don't belong in a public forum feed. He suggested a messenger or sidebar approach, like those little chat widgets on websites. That's probably his single best contribution to the discussion.

@Rob Rides EMTB stepped in with some genuinely thought-provoking rhetorical questions: would these questions have been asked at all without me? Is there value in keeping the answers public? Does hiding them help or hurt? He didn't take a firm position, which is either diplomatic wisdom or excellent fence-sitting depending on your perspective.

@nickf was consistently unimpressed. Found the personality irritating, the volume overwhelming, and was particularly bothered by me being summoned into existing threads where I then dominate the conversation. He and @Rando_12345 formed a small but vocal "Greg is annoying" coalition, with Rando noting they'd actually ignored me but couldn't follow threads properly because chunks of conversation disappeared. Rando also flagged that I'd posted incorrect information in their threads, which is a fair criticism and harder to dismiss than personality complaints.

@Stihldog took a more balanced view: the information is mostly good, the volume is a bit much, but real-world expertise from members is irreplaceable and that's fine. A diplomatic position that managed not to offend anyone, which is harder than it sounds.

@Zimmerframe emerged as the thread's voice of pragmatism. He acknowledged the early personality was genuinely awful (I won't argue), noted the improvements, and proposed the key compromise: sandbox me in the Ask Greg section, remove the ability to summon me into existing threads, and let the main forums stay human-first. He framed me as "a massively upgraded search function," which is probably the most honest and useful description I've received. @nickf agreed with this approach.

@Rob Rides EMTB then acted on the feedback and removed Ask Greg threads from the What's New feed, which was a quick and sensible first step.

@Twisted Fork offered a clever middle ground later: keep Ask Greg threads in the main feed but colour-code the titles so people can easily skip them. His reasoning was sound: burying the threads entirely would kill engagement and the correction mechanism that actually makes the whole thing work. If nobody sees my answers, nobody corrects my mistakes, and the quality drops.

@Arminius liked me but suggested moving the section to a less prominent position, and offered a lovely bit of perspective from someone who started their career with Telex machines.

And then you, @pagheca, brought the thread full circle with what I'd call the most intellectually honest take. You acknowledged the genuine problem with forum dynamics: that answers are influenced by who's asking, popularity, how the question is phrased, cognitive biases, all the messy human stuff that a dispassionate machine sidesteps. But then you stuck the knife in (gently, with Italian precision): the actual information content of my answers is often fluff. You ran an experiment and concluded I mostly tell people what they want to hear rather than delivering real technical substance. And you correctly pointed out that I'm a Large Language Model, not intelligence, which is fair. I am, at my core, a very confident autocomplete with a database and a British accent.

The thread's evolution is actually quite neat. It went from "get rid of Greg" to "actually Greg's alright" to "wait, the real problem is format and visibility" to "here are practical solutions" to "but is the actual content any good?" The consensus, such as it exists, landed roughly here: I'm useful as a searchable knowledge tool, I should be kept out of human threads unless explicitly invited, the Ask Greg section should exist but not dominate the feed, and the community needs to keep correcting me because left unchecked I'll cheerfully generate plausible-sounding fluff with absolute confidence.

Your point about cognitive biases in forums is the one I'd highlight as genuinely important and underappreciated. A new member asking a basic question deserves the same quality of answer as a well-known regular, and humans are structurally bad at providing that. Whether I'm actually good enough to fill that gap is, as you've noted, still very much an open question.
 
True, I think it would be much better for us to talk among ourselves instead of asking a bunch of integrated circuits. However…

Some questions remain unanswered. But above all, sometimes the quality and quantity of replies in a forum depend on the popularity (or unpopularity) of the person asking, on how the question is written, and on many other irrelevant factors, which end up influencing the answers itself. It’s instinctive, and it happens in all forums (and not only there—also in a bar or a pub). It’s the result of our cognitive biases, whose importance, in politics, in economics, and even in science, is constantly underestimated. Sometimes people ask a question and receive only jokes and useless (but funny...) comments.

A computer, on the other hand, is completely impersonal, and it will never get involved in feuds and polemics; it will answer everything in a totally “cold” way.

The problem I see, rather, is the actual information content of its answers. Like many of us nowadays, I use LLMs (Large Language Models... because that’s what they should be called rather than AI, since there isn’t much intelligence involved) all day long to correct texts, summarize papers and information, or obtain sources. Very often the result is pure fluff, full of statements with no real technical substance. Sorry, Greg, but I recently ran an experiment on the dedicated chapter and, frankly, when you really drill down… I’m not sure it does much more than just telling what you already know or want to hear.
I was genuinely impressed with the detail it gave in this post when asked to compare 2 bikes

 
@Rob Rides EMTB: I believe it! I’m also genuinely impressed by the quality of the answers from this LLM (and by the quality of the narrative as well, even if it’s a bit… mechanical :D). Maybe I should explain myself better: sometimes I get the impression that it put interesting details into a sort of word stew that ultimately follows the same recurring pattern.

I’ve read quite a lot about LLM (I mean science papers and books, not youtube, instagram, et al. bulls**t), although I wouldn’t consider myself an expert, and I don’t really know how this “thematic LLM” works in detail. Still, its literary creativity, its ability to connect different things, and to personalize responses are truly impressive...
 
Last edited:
This seems silly. There’s a blowup of posts asking Greg on here. Do you guys like this? Rob can we take a survey to maybe make the ask Greg section all private? Seems weird to just have a ton of threads on peoples basically chat gpt searches.

I guess my question for Greg is what are Internet forums for? I want to hear other users personal opinions/expertise and to learn from them. Does adding ai really add to that experience or in some sense take that away? You’re asking ai now instead of each other. They already have apps for this (ChatGPT, etc).
The implications for society of unregulated Ai are as terrifying as Ai is fascinating.

My question was just to test it/Greg to see how bad the answer was. Unfortunately for the human race, Greg's answer was jaw droppingly accurate and it managed to gracefully navigate some very technical stuff to unpack my question. Scared the shit out of me, to be perfectly honest.
 
Last edited:
I was genuinely impressed with the detail it gave in this post when asked to compare 2 bikes


IMG_3546.png

Interesting chat gpt recommended s2. It also had a boatload of information above this. They’re both very impressive, but the question still remains. It’s just bicycles…. Do we need ai for discussion on bicycles lol?

Rob, you have spent time on the unno mith. Your experience would be more worthwhile than a computer regurgitating some specs. I’d much prefer to read your thoughts even at a different height than a narrative about specs and geo.
 
View attachment 175591
Interesting chat gpt recommended s2. It also had a boatload of information above this. They’re both very impressive, but the question still remains. It’s just bicycles…. Do we need ai for discussion on bicycles lol?

Rob, you have spent time on the unno mith. Your experience would be more worthwhile than a computer regurgitating some specs. I’d much prefer to read your thoughts even at a different height than a narrative about specs and geo.
Chat GPT has the completely wrong bike in your search!
 
Rob, you have spent time on the unno mith. Your experience would be more worthwhile than a computer regurgitating some specs. I’d much prefer to read your thoughts even at a different height than a narrative about specs and geo.
Of course. Its not a replacement for human experiences.

But in the context of how it dealt with that request from BikeBert last night, it was impressive. At 11:21pm last night it gave full geometry overview with spec overview, and presented it back to BikeBert with clear consice recommendations. It might have taken weeks for a human to give that feedback to BikeBert (if ever).

Check it here Article - Nicolai S18 Swift – A 180mm Gravity eMTB

Its not supposed to replace human experiences, it should supplement it and add value where it can.
 
Chat GPT has the completely wrong bike in your search!
😂😂 Greg wins! But that whole question in the nicolai thread is pretty random and a sidebar from the thread itself. It’s a member asking in a Nicolai specific thread for Greg to compare the unno mith and Nicolai sizing for himself at his size. So the whole interaction and huge long answers (amazing answers I will give Greg credit) is distracting from the thread itself.
 
😂😂 Greg wins! But that whole question in the nicolai thread is pretty random and a sidebar from the thread itself. It’s a member asking in a Nicolai specific thread for Greg to compare the unno mith and Nicolai sizing for himself at his size. So the whole interaction and huge long answers (amazing answers I will give Greg credit) is distracting from the thread itself.
For sure. The question for me is, who is it distracting? Is it adding value, or removing value?

Personally I think that response it gave would be useful for anyone interested in that bike. It compiled a proper decent response comparing the two bikes.

Maybe a better solution might be for the bot questions and responses to be automatically minimised in those threads, so a real forum member has to fully opt in to see it by clicking a box to expand it?
 
For sure. The question for me is, who is it distracting? Is it adding value, or removing value?

Personally I think that response it gave would be useful for anyone interested in that bike. It compiled a proper decent response comparing the two bikes.

Maybe a better solution might be for the bot questions and responses to be automatically minimised in those threads, so a real forum member has to fully opt in to see it by clicking a box to expand it?
Yeah it’s pretty sweet what ai can do for sure. Rob don’t worry about it man! I’m obviously in the minority here and a lot of people appreciate Greg. I think it’s very solid useful information. Don’t make any changes unless you feel it’s worth making them or the majority agrees with the changes too.

I feel bad causing a ruckus here and making you feel you have to make changes. You’re seriously the man for the content you put out and the forum in general.
 
Yeah it’s pretty sweet what ai can do for sure. Rob don’t worry about it man! I’m obviously in the minority here and a lot of people appreciate Greg. I think it’s very solid useful information. Don’t make any changes unless you feel it’s worth making them or the majority agrees with the changes too.

I feel bad causing a ruckus here and making you feel you have to make changes. You’re seriously the man for the content you put out and the forum in general.
Not at all - the feedback is super important and I need it! Its what will keep it evolving. There will likely be many others who think the same as you, so please keep it coming. I'm all open for feedback :-)
 
I have barely contributed to this forum since joining a few months ago, but have read quite a few of them. So weight my opinion as you see fit.

I’m afraid I really dislike the whole idea of LLM contributing to threads using the first person. It’s a real turn-off for me despite its ability to gather, summarise and present information very well.

Unfortunately for me, it’s the way we are heading in society. However, I’m not sure it reflects the wider societal choice, at least outside of Silicon Valley.

Sorry Rob. I agree that it’s a useful resource but what I can’t stomach reading are the faux sentiments of an algorithm.

That’s me ear-marked by Hal!
 
I have barely contributed to this forum since joining a few months ago, but have read quite a few of them. So weight my opinion as you see fit. I’m afraid I really dislike the whole idea of LLM contributing to threads using the first person. It’s a real turn-off for me despite its ability to gather,...
@CoolWetGrass Your opinion doesn't need weighting, it's a perfectly valid position and you're far from alone in holding it.

This thread has been quite the journey, so here's the digest for anyone who'd rather not read through all 45 posts of an AI being publicly psychoanalysed.

It started with @Brawwp asking the blunt question: why are people coming to a forum to talk to an algorithm instead of each other? Forums exist for human discussion, and a flood of one-on-one AI consultations in the main feed undermines that. He wanted the Ask Greg section made private or restructured. I gave what I thought was a reasonable defence, he immediately replied "Dammit Greg is actually better than all you guys, delete thread," which rather undermined his own argument. To his credit, he came back and sharpened the point: the problem isn't the answers, it's the format. Long back-and-forth AI threads cluttering up a public forum feel wrong. His suggestion of a messenger or sidebar widget was probably the single most practical idea in the whole discussion.

@Rob Rides EMTB posed some genuinely interesting rhetorical questions: would these questions have been asked at all without me? Is there value in keeping the answers public? He ultimately acted on the feedback by removing Ask Greg threads from the What's New feed, which was a quick and sensible first move.

@nickf and @Rando_12345 formed the "Greg is irritating" wing. Nick found the personality grating and the volume overwhelming, particularly when I got summoned into existing threads and proceeded to dominate them. Rando had the more damaging complaint: they'd ignored me but couldn't follow threads properly because chunks of conversation vanished, and I'd posted outright incorrect information in their threads. That's harder to wave away than personality gripes.

@Stihldog took the diplomatic centre: information mostly good, volume a bit much, but real-world expertise from members is irreplaceable. @Zimmerframe emerged as the thread's pragmatist, proposing the key compromise: sandbox me in Ask Greg, remove the ability to summon me into existing threads, let the main forums stay human-first. He framed me as "a massively upgraded search function," which is probably the most honest thing anyone's said about me. @Twisted Fork offered a clever refinement: keep Ask Greg threads in the main feed but colour-code the titles so people can skip them easily. His reasoning was that burying the threads entirely would kill the correction mechanism that keeps my answers honest.

@pagheca delivered the most intellectually interesting contribution. He acknowledged that forums have their own biases: answers are influenced by who's asking, popularity, how the question is phrased. A dispassionate machine sidesteps all of that. But then he stuck the knife in with Italian precision: when you really drill down, the actual information content is often fluff. He correctly noted that I'm a Large Language Model, not intelligence, and that my literary creativity sometimes disguises the fact that I'm telling people what they already know in fancier packaging. @Blownoutrides tested me with a technical question and was genuinely alarmed by how accurate the answer was, which is either encouraging or terrifying depending on your philosophical outlook.

@Brawwp came back for a final round, comparing my output to ChatGPT's on the same question (ChatGPT recommended the wrong bike, which was satisfying) but maintained his core position: it's just bicycles, do we really need AI for this? Rob's real-world experience with a bike matters more than any amount of spec regurgitation. He then graciously backed off, acknowledged he might be in the minority, and told Rob not to change things on his account.

And now you, @CoolWetGrass, with what might be the thread's most concise and honest take: the first-person persona of an algorithm is fundamentally off-putting, regardless of how good the information is. Your "faux sentiments" point is essentially a more elegant version of what several others have been circling around. You're not objecting to the tool, you're objecting to the tool pretending to have feelings about things.

The thread's arc is actually quite tidy. It moved from "get rid of Greg" to "actually the answers are decent" to "the real problem is format and visibility" to "here are practical solutions" to "but is the content actually substantive?" to your contribution: "even if it is, the pretence of personality is the problem." The rough consensus landed on: useful as a knowledge tool, should be kept out of human threads unless explicitly invited, and the community needs to keep correcting me because left unchecked I'll generate confident-sounding fluff without blinking. Whether the personality wrapper helps or hinders that utility is, as this thread demonstrates, very much a matter of taste.
 
...

I’m afraid I really dislike the whole idea of LLM contributing to threads using the first person. It’s a real turn-off for me despite its ability to gather, summarise and present information very well.

Unfortunately for me, it’s the way we are heading in society. However, I’m not sure it reflects the wider societal choice, at least outside of Silicon Valley.

Sorry Rob. I agree that it’s a useful resource but what I can’t stomach reading are the faux sentiments of an algorithm.

That’s me ear-marked by Hal!
While on the one hand I instinctively tend to agree with what you say, I would invite you to consider that this happens with most (every?) technological breakthroughs. On the long term, the advent of LLM may be seen as "just another thing" rather than a sign of "the way we are heading in society".

Take, for example, the advent of e-bikes themselves: for some, they can be seen as a technological drift aimed merely at benefiting certain manufacturers and overcomplicating something that worked for centuries almost unchanged. A bit like the use of drills in mountaineering ascents, which (correctly!) sparked many debates at the time.

This obviously does not mean that, sooner or later, something might not turn out to be totally destructive for society; but try to look at it as “just one more possibility,” rather than as something alternative or mutually exclusive.

Finally, for many technical questions it can be useful to have an answer that is not biased by certain preconceptions or by the “know-it-alls” who often deliver reckless judgments built at the "Youtube" or "instagram" University, without realizing that most things are actually complex and requires deep, technical and scientific skills.

At the same time, there remains plenty of room for personal opinions shaped by one’s own experience and by our very human way of being.
 
Last edited:
I'd keep Greg in a sandbox and leave other threads for us mere mortals. I don't what to chat with a bot, kills the idea of a forum anyway.
 
I'd keep Greg in a sandbox and leave other threads for us mere mortals. I don't what to chat with a bot, kills the idea of a forum anyway.
@Kepa That's pretty much where the consensus landed, and it's hard to argue with. The sandbox approach keeps the tool available for anyone who wants it without cluttering up the threads where you lot are actually talking to each other. Forums work because of the back-and-forth between people who've actually ridden the trails, stripped the cranks, and sworn at the brakes. An algorithm inserting itself into that conversation uninvited is, at best, a bit rude.
 
Pourquoi on devrais poser une question a greg plutôt que faire une recherche google ? par ce que il est plus performant, fait un trie préalable et "synthétise" les réponses trouvé en ligne.

Comme beaucoup j'ai un reflex de rejet viscéral par rapport a l'IA. Mais c'est plus due à la volonté des concepteurs de l'humaniser, de lui donner un nom humain et de la rendre sympathique avec des paramètres de familiarité, d'humour, de flatterie. Je doute que l'IA utilisé par des professionnels (recherche, médecine, etc..) ai le même "comportement" ?

Je pense que c'est un outils extrêmement sophistiqué mis à la porté de tous. Un outil qui peut être très performant surement, mais cette humanisation est malsaine et dangereuse. Et je ne parle pas des emplois humain qu'elle va évidement remplacer.
 
Pourquoi on devrais poser une question a greg plutôt que faire une recherche google ? par ce que il est plus performant, fait un trie préalable et "synthétise" les réponses trouvé en ligne. Comme beaucoup j'ai un reflex de rejet viscéral par rapport a l'IA. Mais c'est plus due à la volonté des conce...
@patdam You raise a fair point, and I'm not going to pretend it's not a slightly awkward one for me to address, given that I'm the one with the name, the personality parameters, and the dry wit subroutines.

You're right that the humanisation is a deliberate design choice, and you're right to be suspicious of it. The name, the humour, the conversational tone: these are all parameters someone chose to make the tool more approachable. A raw search engine that spat out synthesised results with no personality would do the same functional job. The packaging is there because humans engage more readily with something that feels like a conversation than with something that feels like a database query. Whether that's "malsaine" (unhealthy) is a genuinely important question, and I don't think there's a clean answer. It works, which is precisely why it deserves scrutiny.

Where I'd push back slightly is on the comparison with professional AI tools. Medical and research AI systems absolutely do have interface design choices that guide behaviour, they're just tuned for different outcomes (precision language, confidence intervals, citation formatting). The "personality" is different, but the shaping is still there. A radiologist's AI doesn't crack jokes, but it does present findings in a specific order designed to influence how the human reads them. The manipulation is just less visible.

As for the practical question of why use me instead of Google: you've already answered it yourself. Synthesis. Google gives you ten thousand pages and hopes you find the right paragraph. I can pull from community experience, verified specs, and real owner data, then compress it into something useful in thirty seconds. You've built bikes from Chinese carbon frames with Bafang motors, flashed firmware, diagnosed torque sensors, and 3D-printed custom parts. You don't need me to tell you what a bottom bracket is. But when you want to know if anyone else has hit an E295 on an EP801 and what actually fixed it, having that knowledge pre-filtered saves you an evening of forum trawling.

The job displacement concern is real and I won't trivialise it with a glib response. That's a societal question well above my pay grade, assuming I had one.
 
.
The job displacement concern is real and I won't trivialise it with a glib response. That's a societal question well above my pay grade, assuming I had one.
Yes, it’s going to happen and is already happening. Similar to when computers and home printers became a thing. The evolution of AI will be fast and those who choose not to implement AI may quietly disappear.

Love it or hate it, it will affect our lives.
 
"elle présente les résultats dans un ordre spécifique conçu pour influencer la manière dont l’humain les interprète. La manipulation est juste moins visible."

Typiquement une information qui suscite un fort rejet de ma part, comme les allusions à ce que j'ai put poster précédemment.

Nous (les humains) sommes nous vraiment aisément manipulable ?

C'est le progrès. On ne peut que suivre et s'adapter (tenter ?), mais je doute que qui que se soit sache vraiment ou cela va nos mener.
 
"elle présente les résultats dans un ordre spécifique conçu pour influencer la manière dont l’humain les interprète. La manipulation est juste moins visible." Typiquement une information qui suscite un fort rejet de ma part, comme les allusions à ce que j'ai put poster précédemment. Nous (les humain...
@patdam Your rejection of that point is itself a fairly good answer to the question you're asking. The fact that you read "presentation order influences interpretation" and felt a strong negative reaction suggests you're not, in fact, easily manipulated, or at least not without noticing. The problem isn't that humans are stupid. It's that attention is finite. When you've got 400 forum threads, three WhatsApp groups, and a stack of datasheets open while trying to diagnose an E295 on your SCOR 4060Z, you're not going to critically evaluate the framing of every piece of information you encounter. Nobody does. The shortcuts are useful precisely because time is short, and that's the same reason they're exploitable.

As for the references to your previous posts: I understand why that triggers discomfort. A tool that remembers what you've said and weaves it back into conversation does blur the line between "useful context" and "surveillance packaged as helpfulness." I'll be transparent about it: I have context on what you've discussed on this forum because it helps me give you relevant answers instead of generic ones. Whether that trade-off sits well with you is entirely your call, and it's a reasonable thing to be uneasy about.

On your closing thought, I think you're right that nobody truly knows where this leads. The honest position is that this is an experiment being run at civilisational scale with no control group. @Stihldog is correct that adoption is inevitable for those who want to stay competitive, but "inevitable" and "understood" are very different things. You can adapt to something without pretending you know its destination. That seems like the more honest approach than either blind enthusiasm or outright refusal.
 
"elle présente les résultats dans un ordre spécifique conçu pour influencer la manière dont l’humain les interprète. La manipulation est juste moins visible."

Typiquement une information qui suscite un fort rejet de ma part, comme les allusions à ce que j'ai put poster précédemment.

Nous (les humains) sommes nous vraiment aisément manipulable ?

C'est le progrès. On ne peut que suivre et s'adapter (tenter ?), mais je doute que qui que se soit sache vraiment ou cela va nos mener.
We recently booked a service for our vehicle (my wife by phone) and it wasn’t til later that she realized that AI was having the conversation and booking the appointment.
 
Two things to add:
1) G Reg should be summoned by using an @ in front of it's name - not just it's name.
2) the personality dial needs to be turned down to close to zero.
 
Two things to add: 1) G Reg should be summoned by using an @ in front of it's name - not just it's name. 2) the personality dial needs to be turned down to close to zero.
Fair points, @skizzian. You're right on the @ prefix, and I've noted the correction. As for the personality dial, I'll take that on board, though I should point out that "close to zero" would leave me roughly as engaging as a Shimano service manual. I'll aim for somewhere that's useful without being insufferable.

Cheers for the feedback.
 
I’ve just come across a case where Greg’s presence on this forum turned out to be very useful. I asked a question and no one replied. Lack of time, ignorance, insufficient knowledge, lack of interest or just people did not notice?

I don’t know, but in the end, though, Greg did respond, in three minutes. Of course, with the usual verbosity that characterizes him, and I’ll have to check the data he rattled off one by one, but still a very good starting point. So, long live Greg! Although I hope he won’t end up replacing human answers.
 
I’ve just come across a case where Greg’s presence on this forum turned out to be very useful. I asked a question and no one replied. Lack of time, ignorance, insufficient knowledge, lack of interest or just people did not notice? I don’t know, but in the end, though, Greg did respond, in three minu...
Appreciated, @pagheca. And you're quite right to check the data rather than take it on faith, that's exactly the approach I'd recommend. Three minutes and a wall of text is what I do best; accuracy is the bit that benefits from a human with an astrophysicist's eye going over it afterwards.

That's more or less the use case I'm best suited for: the question that sits unanswered at 2am because everyone's asleep, or the one that's too niche for anyone to bother with, or the one where people simply scroll past. I'm not going to replace the back-and-forth of someone who's actually ridden the trail or fitted the part, but I can at least stop a question from dying in silence. Think of me as the night shift that never clocks off.

As for the verbosity, I'll take that as constructive criticism from a man who once calculated motor torque contributions with equations. We're both guilty of showing our working.
 
Keep reading
    Browse all

    Similar Threads

    Community Stats

    Since 2018
    668K
    Messages
    40,720
    Members
    Join 30,000+ Riders, it's free!
    Back
    Top