Ai - what have you used it for?

[Laughing]

Generative AI or more "traditional" AI / ML ?

Obviously alot of hype is around GenAI but there's plenty of use cases where people are dying to use GenAI because of the hype, but it's equally solved by a standard predictive ML model or something similar.

I work a lot with some of the biggest tech companies in the world (Meta, Amazon etc) and they're still struggling to get beyond the hype an identify proper use cases with real business value even though they have their own MML's an dmanaged services - very much a case of build it and they will come.

I tend to disagree with generative AI. there may not be the revenue streams that the technology deserves, but it can be used very effectively for classification problems in NLP.

I built just this use case for news interntional to classify their backlog of news articles, hundreds of millions of them.

Other use cases for generative AI are for code generation and support though Chat GPT is awful at this, but Cody is very good, but it is trained specifically for code generation, so you would expect it to perform better.

What it can't do is classify things such as medical scans, for which traditional ML is used.

The hype is largely because it is accessible to the masses who want to generate pictures or song lyrics or poems. I guess they are considered impressive because it appears creative.

Open AI, which is the API behind Chat GPT is very impressive as are the organization, whom I have worked with. Chat GPT not so much, but it isn;t intended to be.

 

[Laughing]

The vision system probably is AI

You are right it is. In it's simplest form.

 

[The_Lizards_Jumpers]

I tend to disagree with generative AI. there may not be the revenue streams that the technology deserves, but it can be used very effectively for classification problems in NLP.

I built just this use case for news interntional to classify their backlog of news articles, hundreds of millions of them.

Other use cases for generative AI are for code generation and support though Chat GPT is awful at this, but Cody is very good, but it is trained specifically for code generation, so you would expect it to perform better.

What it can't do is classify things such as medical scans, for which traditional ML is used.

The hype is largely because it is accessible to the masses who want to generate pictures or song lyrics or poems. I guess they are considered impressive because it appears creative.

Open AI, which is the API behind Chat GPT is very impressive as are the organization, whom I have worked with. Chat GPT not so much, but it isn;t intended to be.

No one really knows the barriers of what is possible, there's some great stuff under development at the moment but articulating business value to it and managing the processes and operating model around it is where it's fail or succeed.

 

[Boro_Interlinked]

I used it to play a very cheeky game of thermo nuclear war over dial up modem. It kept finding me on the comms net so we agreed an exhaustive game of tic-tac-toe as a decider.. That's still playing out .. Joshua

 

[Hap]

I had a look at Chat GPT for song lyric inspiration. Yeah, it just spat a song out at me. I didn‘t think much of it tbh.

 

[Ingleby_Flash]

The vision system probably is AI

Maybe, but we were using vision systems 25 years ago, not sure AI was a commercial thing then.

 

[Philly-Red]

On-demand, customised erotica starring myself and any female of my choosing whenever the mood takes me.

 

[Boro_Interlinked]

On-demand, customised erotica starring myself and any female of my choosing whenever the mood takes me.

This is just Weird Science .. you not breaking new ground here

 

[Ingleby_Flash]

On-demand, customised erotica starring myself and any female of my choosing whenever the mood takes me.

Aye, that was the seventies like

 

[Jonathan]

To make daft pictures. Like these ones of Pennywise the Clown looking fed up at a Boro game

 

[Redwurzel]

Used it for my winning lottery ticket numbers.

 

[BoroMart]

I'm looking at the AI app integrated into office 365, teams and the rest of the microsoft suite. It's pretty damned cool. You invite it to your meetings, and it scrapes your emails builds agendas, suggests pre-reads and stakeholders. It then takes notes during your meetings, assigns tasks to people, sends out meeting notes. It can also pre-generate responses to emails for you, you can ask it to give a super formal email response, or "sound like me". It's really cool in supporting your powerpoint building too, and can take an image and manipulate to fit your needs.

 

[Chutney]

I tend to disagree with generative AI. there may not be the revenue streams that the technology deserves, but it can be used very effectively for classification problems in NLP.

I built just this use case for news interntional to classify their backlog of news articles, hundreds of millions of them.

Other use cases for generative AI are for code generation and support though Chat GPT is awful at this, but Cody is very good, but it is trained specifically for code generation, so you would expect it to perform better.

What it can't do is classify things such as medical scans, for which traditional ML is used.

The hype is largely because it is accessible to the masses who want to generate pictures or song lyrics or poems. I guess they are considered impressive because it appears creative.

Open AI, which is the API behind Chat GPT is very impressive as are the organization, whom I have worked with. Chat GPT not so much, but it isn;t intended to be.

At my place we've developed routines that use NLP to convert free text in microbiology results into a queryable structured form. Not sure exactly what tech is used, but most is delivered via Python libraries so it's not exactly reliant on the latest AI for that sort of thing. Newest tool isn't always the right one...