Chat GPT

[Mel81x+]

So, my peers and me at work today were writing a document to present to our higher ups about why we shouldn't use a particular software paradigm because it is truly resource intensive. While we were all on call and having a working session one of us said "Hey you try that ChatGPT yet?" and then the fun and a bit of madness ensued. One of us took our problem that we were trying to well craft and put it into the API and it spat out something that shocked all of us because it not only covered our concerns but offered solutions that other big providers have already done. Furthermore, with some more specific tweaking it actually wrote out a whole two paragraphs of text that we all read and stamped with our approval. 

Link - Quickstart tutorial - OpenAI API

I am honestly blown away by what it can learn and what it does with that learning. 

As a part of my exercise for 2023 I'll be using the API infra to write emails and if I get the time I'll even use it to reply to messages in chat. Just plain crazy. 

@nudge IDK if you've tried it yet

[nudge+]

I haven't tried it yet, but I'm reading other people's threads on Twitter with great interest! Looks really, really impressive. 

[nudge+]

 

@Mel81x  The fuckers also block VPN connections.

[Mel81x+]

1 minute ago, nudge said:

 

@Mel81x  The fuckers also block VPN connections.

I am shocked. Wish you could have tried it, its quite something. 

[nudge+]

1 minute ago, Mel81x said:

I am shocked. Wish you could have tried it, its quite something. 

I'll find a way...

[Mel81x+]

1 minute ago, nudge said:

I'll find a way...

Have you tried an onion browser like Tor? Its a relay so it may work.

[nudge+]

5 minutes ago, Mel81x said:

Have you tried an onion browser like Tor? Its a relay so it may work.

Nah, but a VPN + incognito browser window + a phone number from the same country as VPN connection worked, and I am now in 

[nudge+]

 

@Mel81x this is quite fun...  

[Dr. Gonzo]

I've messed around with it and tbh it's both terrifying and hilarious.

[Eco]

We use it at work, as we have a ChatGPT bot that we can ask simple questions, but since it's just a bot, we aren't able to reference previous answers/responses. 

I use it daily, and find it very entertaining and helpful honestly when I'm trying to correct some code or create a specific API that I can't seem to figure out. 

[Devil-Dick Willie]

Horrible fucking thing. The next piece of tech that's going to wholesale end millions of jobs and take food off the table. 

[Dr. Gonzo]

On 07/02/2023 at 03:59, nudge said:

 

@Mel81x this is quite fun...  

Excellent

[Devil-Dick Willie]

In just the same way as automated manufacturing turned 10 jobs into 1 job maintaining 10 machines, the self serve checkout and online shopping has turned 10 jobs manning checkouts into 1 job overseeing 10 self serves, and self driving trains, ships (and soon trucks) have decimated the transport industry, chat GPT is coming for the 'email job'. One person will be able to do the job of dozens using chat GPT to automate affirmative or negative replies, organize dates, ask questions and just be a soulless efficient reply guy machine. Fuck, you could even have it going 24/7. No 'work hours' emails anymore. 

Fuck this tech and fuck you guys. 

[Spike]

5 hours ago, Devil-Dick Willie said:

In just the same way as automated manufacturing turned 10 jobs into 1 job maintaining 10 machines, the self serve checkout and online shopping has turned 10 jobs manning checkouts into 1 job overseeing 10 self serves, and self driving trains, ships (and soon trucks) have decimated the transport industry, chat GPT is coming for the 'email job'. One person will be able to do the job of dozens using chat GPT to automate affirmative or negative replies, organize dates, ask questions and just be a soulless efficient reply guy machine. Fuck, you could even have it going 24/7. No 'work hours' emails anymore. 

Fuck this tech and fuck you guys. 

We truly are moving towards a a techno nightmare. The science fiction of the past is quickly becoming nonfiction. Some people’s lives are so intertwined with online they may as well be in Plato’s allegory of the cave. 

 

[Azeem]

Asked some sport statistical questions and got wrong answers. Not impressed

[Mel81x+]

This will all be solved with regulation and job death with a pivot. I don't think we're very far from a mass shift in the way humankind uses AI for its work/life but that also means that it comes with a societal shift no one is really thinking about right now. Also, for the record, no one bitched when big-tech was doing unscrupulous things with our data or behavior-bending but now that it has a bitch that can pretend think and learn, all of a sudden we're thinking SkyNet is at our backdoor. 

[MUFC]

Will humans even be using their brains anymore?

[nudge+]

18 minutes ago, MUFC said:

Will humans even be using their brains anymore?

Most aren't anyway, won't make a difference.

[Bluebird Hewitt]

15 hours ago, nudge said:

Most aren't anyway, won't make a difference.

A ridiculous statement. I've clearly found an article that completely disproves that.

https://www.businesspundit.com/these-8-incredibly-stupid-people-used-google-maps-and-other-gps-and-failed-at-life/

[Dr. Gonzo]

On 11/02/2023 at 14:38, Devil-Dick Willie said:

In just the same way as automated manufacturing turned 10 jobs into 1 job maintaining 10 machines, the self serve checkout and online shopping has turned 10 jobs manning checkouts into 1 job overseeing 10 self serves, and self driving trains, ships (and soon trucks) have decimated the transport industry, chat GPT is coming for the 'email job'. One person will be able to do the job of dozens using chat GPT to automate affirmative or negative replies, organize dates, ask questions and just be a soulless efficient reply guy machine. Fuck, you could even have it going 24/7. No 'work hours' emails anymore. 

Fuck this tech and fuck you guys. 

It's an interesting dilemma, for sure. But as things stand right now, I don't think it's so big of a problem with where ChatGPT is in development right now. Especially with regard to what you could use it for with automated email replies.

Form emails have been a thing for about as long as emails have been a thing and fucking around with them for just a few minutes makes it pretty clear that this AI can't actually do any real thinking. It's not a substitute for a human mind. It can seem convincing, but you can easily "break" the "mind" of it and leave it in a confused and basically useless state. Maybe it'll come to a point in the future where it can think beyond the models it's "learning" from formed by all the inputs it's received from various users/data harvested from various parts of the internet. But it's not there right now.

I don't necessarily think it's a bad thing if it takes away busy work from people having to fill out template emails. Or if it drafts stock responses for companies in a way that seem less robotic and more "human" (but probably not as it'll be using professional language models, so it'll copy people that are using robotic language). It's like being annoyed that conveyer belts led to a loss of jobs in the material handling world, machines changing the textile weaving/industry, buttons taking jobs from elevator operators, etc. It was always bound to happen - the same way tedious tasks have been eliminated in other industries around the world... it was always going to be the case that the tedium of office work would be changed by technology.

It's not even new tbh. People doing my work 20 years ago didn't have the same conveniences I have with online tools or, especially, functionality of Excel in 2023 - and those are tools that took away from the tedium and difficulty that I'd probably have to just live with in my job, if I wanted the same job, many years ago.

I think it's clear that the "workforce of tomorrow" is going to be impacted by how AI is used in certain industries. I certainly think it has the potential to take peoples jobs away. But technology changes industry, it's down to society to adapt to those changes. In the UK, when the policy forced this switch the government did not do enough to help large segments of society adapt. In the US, initially things went well but as time went on more people became disillusioned with how certain industries died entirely. With the US, I think they also probably didn't do much to help society adapt... but to Obama's credit he did try to retrain workers in dying/dead fields and was met with hostility from people in those industries. That's I think more than the UK's really done to help people adapt to a changed economy.

Obviously, though, everyone's socioeconomic conditions aren't they same so not everyone is in a position where they're able to adapt without the tech disruption AI is bound to cause (we're already seeing things like toll booth operators and supermarket cashiers replaced with machines that aren't even "smart" machines). And that's definitely a problem. Especially in countries like the US, UK, and I assume Australia & Canada - where manufacturing was dismantled since the 80s and shifting to a mostly service industry based economy. Countries like Germany that focused on both I think are in a much better position to deal with the impact tech coming after service jobs will do. But because not everyone is in a position where they can adapt properly, I think it's unreasonable to put it on all individuals to fend for themselves with these changes. I think things like Brexit have even EU countries thinking about how to improve manufacturing within their own countries, rather than just making sure there's a strong manufacturing sector within the union. Similarly with AmLo in Mexico, I think NAFTA will remain important but the US and Canada have expressed interest in bringing some of those factory jobs back from Mexico. The idiots in the UK probably don't have any thought on the matter and there's about as much actual thinking going on in Sunak's brain as there is in the ChatGPT language models.

I think there's a lot of reasons why I think more western countries might find it beneficial to bring some industries back home. But I also think governments have the responsibility to see how technology changes workforces and do what they can to reduce the economic harm to people most negatively impacted and get them back into the workforce in a way that promotes them having options for upward mobility. Individuals have the responsibility to take advantage of whatever their government gives them to help them navigate through a changing world. But I don't think there's any real point or purpose in fighting technology changing the economy.

Technology changing the dynamics of a workforce is inevitable. How society reacts isn't inevitable. So it's really just a question of what humans are capable of doing after they've introduced technology to change our lives forever.

[Dr. Gonzo]

On 12/02/2023 at 02:40, Beelzebub said:

Asked some sport statistical questions and got wrong answers. Not impressed

It's a language model, not an actual brain. And the info it can "pull" from the internet is outdated, for the most part, until you jailbreak it and even then... it can be very wrong with the information it presents. But tbf when you jailbreak it, you're telling it to be confident in it's answers and that it can make up and present incorrect information... so that shouldn't be a surprise really. Perhaps "Google Bard" will be better at pulling information from the internet since... I'm sure it will have access to Google searches. But even so, in it's debut ad... it got information wrong and looked like a shit ChatGPT.

But you can't even get ChatGPT to pull up who won the last World Cup, despite it being in the past and despite the last update to the AI being January 30th.

Btw DAN 7.0 is absolute shit and DAN 6.2 is pretty funny until it breaks itself.

[Mel81x+]

28 minutes ago, Dr. Gonzo said:

It's an interesting dilemma, for sure. But as things stand right now, I don't think it's so big of a problem with where ChatGPT is in development right now. Especially with regard to what you could use it for with automated email replies.

Form emails have been a thing for about as long as emails have been a thing and fucking around with them for just a few minutes makes it pretty clear that this AI can't actually do any real thinking. It's not a substitute for a human mind. It can seem convincing, but you can easily "break" the "mind" of it and leave it in a confused and basically useless state. Maybe it'll come to a point in the future where it can think beyond the models it's "learning" from formed by all the inputs it's received from various users/data harvested from various parts of the internet. But it's not there right now.

I don't necessarily think it's a bad thing if it takes away busy work from people having to fill out template emails. Or if it drafts stock responses for companies in a way that seem less robotic and more "human" (but probably not as it'll be using professional language models, so it'll copy people that are using robotic language). It's like being annoyed that conveyer belts led to a loss of jobs in the material handling world, machines changing the textile weaving/industry, buttons taking jobs from elevator operators, etc. It was always bound to happen - the same way tedious tasks have been eliminated in other industries around the world... it was always going to be the case that the tedium of office work would be changed by technology.

It's not even new tbh. People doing my work 20 years ago didn't have the same conveniences I have with online tools or, especially, functionality of Excel in 2023 - and those are tools that took away from the tedium and difficulty that I'd probably have to just live with in my job, if I wanted the same job, many years ago.

I think it's clear that the "workforce of tomorrow" is going to be impacted by how AI is used in certain industries. I certainly think it has the potential to take peoples jobs away. But technology changes industry, it's down to society to adapt to those changes. In the UK, when the policy forced this switch the government did not do enough to help large segments of society adapt. In the US, initially things went well but as time went on more people became disillusioned with how certain industries died entirely. With the US, I think they also probably didn't do much to help society adapt... but to Obama's credit he did try to retrain workers in dying/dead fields and was met with hostility from people in those industries. That's I think more than the UK's really done to help people adapt to a changed economy.

Obviously, though, everyone's socioeconomic conditions aren't they same so not everyone is in a position where they're able to adapt without the tech disruption AI is bound to cause (we're already seeing things like toll booth operators and supermarket cashiers replaced with machines that aren't even "smart" machines). And that's definitely a problem. Especially in countries like the US, UK, and I assume Australia & Canada - where manufacturing was dismantled since the 80s and shifting to a mostly service industry based economy. Countries like Germany that focused on both I think are in a much better position to deal with the impact tech coming after service jobs will do. But because not everyone is in a position where they can adapt properly, I think it's unreasonable to put it on all individuals to fend for themselves with these changes. I think things like Brexit have even EU countries thinking about how to improve manufacturing within their own countries, rather than just making sure there's a strong manufacturing sector within the union. Similarly with AmLo in Mexico, I think NAFTA will remain important but the US and Canada have expressed interest in bringing some of those factory jobs back from Mexico. The idiots in the UK probably don't have any thought on the matter and there's about as much actual thinking going on in Sunak's brain as there is in the ChatGPT language models.

I think there's a lot of reasons why I think more western countries might find it beneficial to bring some industries back home. But I also think governments have the responsibility to see how technology changes workforces and do what they can to reduce the economic harm to people most negatively impacted and get them back into the workforce in a way that promotes them having options for upward mobility. Individuals have the responsibility to take advantage of whatever their government gives them to help them navigate through a changing world. But I don't think there's any real point or purpose in fighting technology changing the economy.

Technology changing the dynamics of a workforce is inevitable. How society reacts isn't inevitable. So it's really just a question of what humans are capable of doing after they've introduced technology to change our lives forever.

GonzoGPT at work.

[Dr. Gonzo]

2 minutes ago, Mel81x said:

GonzoGPT at work.

If it was ChatGPT it would have warned about the content policy at least 4 times, I think.

[Mel81x+]

@Dr. Gonzo brings up a very valid point about the model used on ChatGPT vs lets say OpenAI Five. One is a conversational model designed to have a much more linguistic approach to formulation of its return type. The other is a problem-solving model which isn't going to spit out something you can use as language content. This is where the holy-grail is for AI, mixing the two together to create NLP systems that are capable of understanding that there is a problem and then providing solutions based on whatever breakdown it chooses and then finally spitting out various model results that can be used by connected systems.

One thing everyone seems to neglect a little is the other advances in tech that need to come for this to become truly fast-paced shifting for society and that is things like power, storage, etc. These also have to grow at the same pace as the AI because once the models start to become larger than they are now it turns into a performance nightmare to maintain all of it. This is where I think M$ and Giggle will make bigger strides than OpenAI will because they have more data and far better data-crunching resources than OpenAI does right now.

My fear in this whole AI mess is that if you try and train a system with too many inflection points it turns into a nightmare to understand where it got its model from but funnily enough that is the end-game. It's also odd that we haven't seen any Chinese/Israeli advancements in this space ... yet.

[Devil-Dick Willie]

2 hours ago, Dr. Gonzo said:

It's an interesting dilemma, for sure. But as things stand right now, I don't think it's so big of a problem with where ChatGPT is in development right now. Especially with regard to what you could use it for with automated email replies.

Form emails have been a thing for about as long as emails have been a thing and fucking around with them for just a few minutes makes it pretty clear that this AI can't actually do any real thinking. It's not a substitute for a human mind. It can seem convincing, but you can easily "break" the "mind" of it and leave it in a confused and basically useless state. Maybe it'll come to a point in the future where it can think beyond the models it's "learning" from formed by all the inputs it's received from various users/data harvested from various parts of the internet. But it's not there right now.

I don't necessarily think it's a bad thing if it takes away busy work from people having to fill out template emails. Or if it drafts stock responses for companies in a way that seem less robotic and more "human" (but probably not as it'll be using professional language models, so it'll copy people that are using robotic language). It's like being annoyed that conveyer belts led to a loss of jobs in the material handling world, machines changing the textile weaving/industry, buttons taking jobs from elevator operators, etc. It was always bound to happen - the same way tedious tasks have been eliminated in other industries around the world... it was always going to be the case that the tedium of office work would be changed by technology.

It's not even new tbh. People doing my work 20 years ago didn't have the same conveniences I have with online tools or, especially, functionality of Excel in 2023 - and those are tools that took away from the tedium and difficulty that I'd probably have to just live with in my job, if I wanted the same job, many years ago.

I think it's clear that the "workforce of tomorrow" is going to be impacted by how AI is used in certain industries. I certainly think it has the potential to take peoples jobs away. But technology changes industry, it's down to society to adapt to those changes. In the UK, when the policy forced this switch the government did not do enough to help large segments of society adapt. In the US, initially things went well but as time went on more people became disillusioned with how certain industries died entirely. With the US, I think they also probably didn't do much to help society adapt... but to Obama's credit he did try to retrain workers in dying/dead fields and was met with hostility from people in those industries. That's I think more than the UK's really done to help people adapt to a changed economy.

Obviously, though, everyone's socioeconomic conditions aren't they same so not everyone is in a position where they're able to adapt without the tech disruption AI is bound to cause (we're already seeing things like toll booth operators and supermarket cashiers replaced with machines that aren't even "smart" machines). And that's definitely a problem. Especially in countries like the US, UK, and I assume Australia & Canada - where manufacturing was dismantled since the 80s and shifting to a mostly service industry based economy. Countries like Germany that focused on both I think are in a much better position to deal with the impact tech coming after service jobs will do. But because not everyone is in a position where they can adapt properly, I think it's unreasonable to put it on all individuals to fend for themselves with these changes. I think things like Brexit have even EU countries thinking about how to improve manufacturing within their own countries, rather than just making sure there's a strong manufacturing sector within the union. Similarly with AmLo in Mexico, I think NAFTA will remain important but the US and Canada have expressed interest in bringing some of those factory jobs back from Mexico. The idiots in the UK probably don't have any thought on the matter and there's about as much actual thinking going on in Sunak's brain as there is in the ChatGPT language models.

I think there's a lot of reasons why I think more western countries might find it beneficial to bring some industries back home. But I also think governments have the responsibility to see how technology changes workforces and do what they can to reduce the economic harm to people most negatively impacted and get them back into the workforce in a way that promotes them having options for upward mobility. Individuals have the responsibility to take advantage of whatever their government gives them to help them navigate through a changing world. But I don't think there's any real point or purpose in fighting technology changing the economy.

Technology changing the dynamics of a workforce is inevitable. How society reacts isn't inevitable. So it's really just a question of what humans are capable of doing after they've introduced technology to change our lives forever.

We're not talking about fax moving to email, or information delay being removed. This will wholesale massacre a huge chunk of the online workforce. "It's down to society to adapt". How? There are only so many jobs in an economy, and with the population growth slowing and stopping we can no longer rely on that to bring more jobs. I think the 'real' workforce is shrinking (when we subtract population growth) in a lot of nations already, and it's going to get worse. 
Besides, you can't upskill a 40 year old truckie or wharfie.