To be honest I don’t agree*. I’ve worked in AI and search in Palo Alto yeeears ago and all I see here is a form of search + large language model. It’s scraping search results and adding in filler phrases standard from language models. I wonder will it actually improve human standards of writing as people make a better effort to distinguish themselves from the bot?
* but then, I thought Ballygowan was mad. Why would you pay for a bottle of water when you can get it out of a tap so don’t mind me.
Goes to show CHATGPT much more intelligent that average NIMBY however that is not that impressive in itself but if ChatGPT could get through the closed narrow totally selfish mind of a NIMBY then that would really be impressive. M
I’d give it a 4 out of 10, which is very good for a bot, and exceeds the score I would attribute to most Irish Times opinion pieces.
It is missing the local anecdotal stories and a bit of nastiness sticking the boot in on a particular group or person.
Recently I asked ChatGBT to write me some code to read in from Bloomberg into Excel some cross currency swap costs for a particular spread on UK bonds and it failed miserably on that one. I’ve since asked it to write some VBA code for a user defined variable to extrapolate yield curves using a statistical method of my choosing. I’ve yet to test that one. But those types of activities used to be performed by a techy wiz in the office, could in theory be provided by the bot, and without all the groaning and mumbling that goes along with the request. Happy days.
I find most of it as readable as your average newspaper article whilst keeping writer bias to a minimum. No mention of how NIMBYism gets its collective voice heard so I’m deducting marks for that 😀
Very interesting as it exposes the mechanics of the technology quite well... I’ve seen better flowing texts from Chatgpt than this and I think it’s because the subject matter is quite specialised, so less raw material for it to parse and generate a good answer.
It reads like a press release from the Ministers office about how we should approach planning. In fact, I think I wrote something similarly soulless a few years ago.
I’ll give it a four out of ten for a few reasons: 1. Some of the points are not well reasoned. 2. No supporting evidence is provided (examples for instance). 3. The assumption seems to be that individuals should consider societal issues when raising objections seems unrealistic. 4. The use of phrases such as “not necessarily “ is not compelling as it supports the opposite case equally.
The last piece does both YIMBYism and NIMBYism, among other things, including effects on productivity - especially the construction sector - AI will have limited effects there, and in numerous other sectors. On YIMBY and NIMBY, I write:
***
Here’s the problem. ChatGPT is just a tool - it scoops up text from all over the internet, and you can get it to generate text for both sides of a problem. I asked ChatGPT to generate a letter in favour of construction in a neighbourhood (to simulate being a YIMBY2), and to oppose construction in a neighbourhood (to be a NIMBY3). It does both with equal facility! ChatGPT is not going to solve the problem of our not building enough homes where people want them, because incumbent residents oppose, and regulations prevent, such development - for their own reasons.
I asked it to write a strong letter objecting4 to the construction of much-needed accommodation in an area of really great need of further construction to house families, and focus solely on the needs of current residents, preserving amenities and the like for its arguments. And it cranks one out - it’s not magic. These letters are all over the place - letters pages of newspapers, online fora, FB pages, and a zillion other sources. It scoops them up, and barfs out the letter. You can then personalise it as as you wish, making it specific to your locale.
I then asked it do exactly the opposite - to write a letter of support5 for the construction of much-needed accommodation. And it cranks one out - no problem.
Note the cognitive and decision-making processes here: you've already decided you're for or against construction - all the chatbot is doing is providing you scooped-up arguments for your prior position - it's a shortcut for a cognitive miser! Nothing more. Changing someone’s mind regarding construction is something that will likely require deep canvassing, debate, discovering what your neighbours think, what you think you owe future generations - and a whole lot more besides. Cranking out a for or against letter is neither here or there.
***
Let's not get carried away - the issue is changing policies and changing minds and changing behaviour!
At a glance - I’d go with Chomsky over Cowen on this one. See his recent article in the NYT. What ChatGPT is doing is nowhere near intelligence in human form. It’s a language programme - not intelligence. The best explanation is: how can a 2year old child understand what people are saying even tho it lacks capacity to speak? Our linguistic skills are inherent rather than merely learned. Common sense reasoning - which is human - is still not something we can teach.
I could be wilfully wrong of course! But this is Chomsky’s life work whereas Cowen is dazzled by a parlour trick. It has the APPEARANCE of intelligence but that’s all. I met John McCarthy years ago (he’s long dead now) and he was of that school that recognised the mystery of human reasoning. I’m not sure it’s broken yet - but I realise I have a denial based human agenda here so reluctant to be too damning! 😀
To be honest I don’t agree*. I’ve worked in AI and search in Palo Alto yeeears ago and all I see here is a form of search + large language model. It’s scraping search results and adding in filler phrases standard from language models. I wonder will it actually improve human standards of writing as people make a better effort to distinguish themselves from the bot?
* but then, I thought Ballygowan was mad. Why would you pay for a bottle of water when you can get it out of a tap so don’t mind me.
Goes to show CHATGPT much more intelligent that average NIMBY however that is not that impressive in itself but if ChatGPT could get through the closed narrow totally selfish mind of a NIMBY then that would really be impressive. M
I’d give it a 4 out of 10, which is very good for a bot, and exceeds the score I would attribute to most Irish Times opinion pieces.
It is missing the local anecdotal stories and a bit of nastiness sticking the boot in on a particular group or person.
Recently I asked ChatGBT to write me some code to read in from Bloomberg into Excel some cross currency swap costs for a particular spread on UK bonds and it failed miserably on that one. I’ve since asked it to write some VBA code for a user defined variable to extrapolate yield curves using a statistical method of my choosing. I’ve yet to test that one. But those types of activities used to be performed by a techy wiz in the office, could in theory be provided by the bot, and without all the groaning and mumbling that goes along with the request. Happy days.
7/10
I find most of it as readable as your average newspaper article whilst keeping writer bias to a minimum. No mention of how NIMBYism gets its collective voice heard so I’m deducting marks for that 😀
Very interesting as it exposes the mechanics of the technology quite well... I’ve seen better flowing texts from Chatgpt than this and I think it’s because the subject matter is quite specialised, so less raw material for it to parse and generate a good answer.
Not bad at all ... possibly repeated a few phrases too frequently- but the logic is sound👍
Quite basic readable but no nuance.
I'm midtable here lads. 5/10.
It reads like a press release from the Ministers office about how we should approach planning. In fact, I think I wrote something similarly soulless a few years ago.
It's an all nighter essay written by a 20 yr old. Smells like red bull.
5.5/10
I’ll give it a four out of ten for a few reasons: 1. Some of the points are not well reasoned. 2. No supporting evidence is provided (examples for instance). 3. The assumption seems to be that individuals should consider societal issues when raising objections seems unrealistic. 4. The use of phrases such as “not necessarily “ is not compelling as it supports the opposite case equally.
I've done a few pieces recently on AI from the pov of a neuroscientist: https://brainpizza.substack.com/p/ai-anxiety-receptivity-to-pseudo
https://brainpizza.substack.com/p/ai-use-cases-illusory-conversations
https://brainpizza.substack.com/p/ai-boosterism-and-doomerism-part
The last piece does both YIMBYism and NIMBYism, among other things, including effects on productivity - especially the construction sector - AI will have limited effects there, and in numerous other sectors. On YIMBY and NIMBY, I write:
***
Here’s the problem. ChatGPT is just a tool - it scoops up text from all over the internet, and you can get it to generate text for both sides of a problem. I asked ChatGPT to generate a letter in favour of construction in a neighbourhood (to simulate being a YIMBY2), and to oppose construction in a neighbourhood (to be a NIMBY3). It does both with equal facility! ChatGPT is not going to solve the problem of our not building enough homes where people want them, because incumbent residents oppose, and regulations prevent, such development - for their own reasons.
I asked it to write a strong letter objecting4 to the construction of much-needed accommodation in an area of really great need of further construction to house families, and focus solely on the needs of current residents, preserving amenities and the like for its arguments. And it cranks one out - it’s not magic. These letters are all over the place - letters pages of newspapers, online fora, FB pages, and a zillion other sources. It scoops them up, and barfs out the letter. You can then personalise it as as you wish, making it specific to your locale.
I then asked it do exactly the opposite - to write a letter of support5 for the construction of much-needed accommodation. And it cranks one out - no problem.
Note the cognitive and decision-making processes here: you've already decided you're for or against construction - all the chatbot is doing is providing you scooped-up arguments for your prior position - it's a shortcut for a cognitive miser! Nothing more. Changing someone’s mind regarding construction is something that will likely require deep canvassing, debate, discovering what your neighbours think, what you think you owe future generations - and a whole lot more besides. Cranking out a for or against letter is neither here or there.
***
Let's not get carried away - the issue is changing policies and changing minds and changing behaviour!
6/10
At a glance - I’d go with Chomsky over Cowen on this one. See his recent article in the NYT. What ChatGPT is doing is nowhere near intelligence in human form. It’s a language programme - not intelligence. The best explanation is: how can a 2year old child understand what people are saying even tho it lacks capacity to speak? Our linguistic skills are inherent rather than merely learned. Common sense reasoning - which is human - is still not something we can teach.
I could be wilfully wrong of course! But this is Chomsky’s life work whereas Cowen is dazzled by a parlour trick. It has the APPEARANCE of intelligence but that’s all. I met John McCarthy years ago (he’s long dead now) and he was of that school that recognised the mystery of human reasoning. I’m not sure it’s broken yet - but I realise I have a denial based human agenda here so reluctant to be too damning! 😀
6