ChatGPT AI Tool— We all work for robots now

[Captainant 28381]

6 hours ago, MC Fresh Breath said:

I quote that as someone who uses ChatGPT from time to time.  Mainly to help format code or find an answer that I don't have to read through various Stack Oveflow answers on now.  It also handily comes w/o the derision and arguing that can be present on Stack Overflow.  Doesn't always get things right, but it is usually close.  I also know when I do this I'm helping to train it.  Ah well.

You should use a coding specific tool like GitHub copilot or AWS codewhisperer, those LLMs are fine tuned for code and usually give better results than a generalist foundational model

[FartingMonk 3612]

Facebook just announced them going all in on AI.  Those AI servers are power intensive.  

[MC Fresh Breath 4075]

On 3/2/2024 at 8:58 PM, Captainant said:

You should use a coding specific tool like GitHub copilot or AWS codewhisperer, those LLMs are fine tuned for code and usually give better results than a generalist foundational model

yep.  have copilot at work. i still often find the chatgpt engagement a bit better edit: for random tasks.  for sure copilot when i have code in flight is quite a bit more direct.  however i enjoy sometimes 'engaging' with chatgpt for more of the 'why.'

 

Edited March 4 by MC Fresh Breath

[Captainant 28381]

My team at work just got security approval for a new set of in-house built tools that are pretty sweet. If we record a meeting, we can throw the .mp4 at a slack bot and it'll transcribe the call and then author a summary that gives stakeholders in the call, key details, decisions made, and action items for each stakeholder identified.

Saves me a good 15 to 20 minutes of cleaning up my notes from a call, and my customers LOVE it! 

It's not replacing me, but it's making me a hell of a lot more efficient and better documented

[Nice Guy Eddie 13706]

22 minutes ago, Captainant said:

My team at work just got security approval for a new set of in-house built tools that are pretty sweet. If we record a meeting, we can throw the .mp4 at a slack bot and it'll transcribe the call and then author a summary that gives stakeholders in the call, key details, decisions made, and action items for each stakeholder identified.

Saves me a good 15 to 20 minutes of cleaning up my notes from a call, and my customers LOVE it! 

It's not replacing me, but it's making me a hell of a lot more efficient and better documented

This gives you more time to spend on Surly. 

[gmr548 16681]

On 2/21/2024 at 1:27 PM, Goredho said:

https://futurism.com/chatgpt-lost-mind
 

Uh oh.

As The Independent reports, ChatGPT users have spent the last 24 hours or so flocking to social media to share screenshots and anecdotes of bizarre interactions with the OpenAI chatbot — which, well, appears to be losing its mind.

Screenshots show the AI's responses to seemingly normal queries devolving into total gibberish, or simply generating way too much content. In one case highlighted by the Independent, a Redditor shared that the AI — when asked a question about coding, mind you — provided a garrulous and mostly illogical answer that included the statement: "let's keep the line as if AI in the room."

more in the link

ChatLobo/Cruiser/YGIFS?

[Captainant 28381]

43 minutes ago, BeardIP said:

This? https://www.salesforce.com/news/press-releases/2024/02/27/einstein-copilot-news/

Nope, but that looks pretty similar to Amazon's Q or Azure's AI Language where you can integrate it with your knowledge sources (confluence, sharepoint, ServiceNow, etc) and it can help answer contextual questions - but it's tied into that one system. We've got an in-house built system that started as a tech learning project and grew into an nice little internal tool, but under the covers it's just interfaces and API's so new intents and functionality is coming every month or so as more folks contribute to it

Edited March 5 by Captainant

[crash_davis 23465]

Robot is pervy.

 

[atomheartbevo 90543]

Is there anything stopping people from putting out albums that sound like famous singers/bands?

I mean like if you cranked out an album of cover songs from various famous singers/bands, would you be sued, like the song below.

Asking for a friend.

 

[Captainant 28381]

On 1/29/2024 at 2:36 PM, Captainant said:

It's not chatGPT, but is another great example of the "AI' really being offshore humans via a very elaborate Mechanical Turk

https://www.livemint.com/companies/news/amazons-ai-based-just-walk-out-checkout-tech-was-powered-by-1000-indian-workers-manually-11712196827721.html

[956 Worldwide 35516]

AI: Actually Indians 

[Incredulity 11859]

On 3/4/2024 at 1:50 PM, Captainant said:

and better documented

**Surly plantiff's lawyer licks chops.

[bernorange 8385]

Competitor for ChatGPT et al:

https://venice.ai/chat

Info:

https://moneyandstate.com/blog/the-separation-of-mind-and-state

[Indomitable 80]

 

[atomheartbevo 90543]

On 3/2/2024 at 3:41 PM, thunderlounge said:

Really could be a problem in the wrong hands. And it’s already there. I don’t think the tech itself is the scary part, but if people believe the falsehoods created. 

I think we need to show the people who are gullible to this stuff, this video from 1984 because Stewart and Herb knew what was up on Computer Chronicles.

 

[General Specific 3484]

I think we need to show the people who are gullible to this stuff, this video from 1984 because Stewart and Herb knew what was up on Computer Chronicles.
 

+1 for that blast from the past.

And +1 for Lady Ada Lovelace, we’ve been at this AI thing a very very long time.

[BurntEyes 6198]

On 3/4/2024 at 4:13 PM, Nice Guy Eddie said:

This gives you more time to spend on Surly. 

This and reduces he said she said non-sense or horrible interpretation.

The team and AI agree that the client/contact said X, George. You're the only one who thinks they were talking about shrimp the entire time.

[hayden_horn 11401]

the problem with ai is that it is reductive, because it has to be, but it is beng advertised as "generative"

all it's generating is a response indexed from the internet or a knowledge base.

to the effect that it is creating something "new" it is doing so out of used parts, if you will.

as such, it can not actually be generative or creative in any real sense.

[tbone_ 14406]

the problem with ai is that it is reductive, because it has to be, but it is beng advertised as "generative"
all it's generating is a response indexed from the internet or a knowledge base.
to the effect that it is creating something "new" it is doing so out of used parts, if you will.
as such, it can not actually be generative or creative in any real sense.

Sounds like something a robot would say

[atomheartbevo 90543]

9 hours ago, General Specific said:

+1 for that blast from the past.

And +1 for Lady Ada Lovelace, we’ve been at this AI thing a very very long time.

There's like around 500 or so episodes up on Youtube - Stewart had kept all of the archives of the original studio reels, so even though they weren't quite 35mm or even 16mm, they still look pretty good for what was basically a weekly PBS show starting in the 80s.  He released them all to the Internet Archive and YouTube.

And the fascinating thing about that show is that the SRI International that the co-host Herb is a part of, as are some of the interviewees,  is Stanford Research Institute, which had an AI lab in 1966.  At the point that show taped, that lab had been humming along for 18 years.

I've been working my way through the shows as they pop up (it's on my bucket list to finish all of them) and AI got a lot of mention in the earlier shows, and then in the later 80s it stopped being talked about, which mirrors the industry as a whole. I haven't watched a lot of 90s episodes, so I don't know if it gets a mention that much (I'm focusing on the Gary Kildall era first).

Edited July 23 by atomheartbevo

[immamac 44199]

In other news meta released llama 3.1 open source and it is the same shit. 

[BurntEyes 6198]

Rich Heimann wrote a really good book called Doing AI.

It's great for the layidiots like me to understand the weird world of AI, both in business and in academics.  

Edited July 23 by BurntEyes

[speed817 19844]

🤣 🤣 🤣

[totallynotabuttpirate 5477]

"late-night boba-run" is one of the sickest burns ever. 

[Captain Ron 13095]

[pacman 7485]

 

[Parliament 14556]

I see commercials from these companies (Google mostly I think) advertising AI as your "new friend."  Like one that shows how AI can coach you through an ice bath.  

Is there REALLY a market for fake friends?  Am I out of touch because that's sad, scary and stupid.

[NoName 17105]

4 hours ago, Parliament said:

I see commercials from these companies (Google mostly I think) advertising AI as your "new friend."  Like one that shows how AI can coach you through an ice bath.  

Is there REALLY a market for fake friends?  Am I out of touch because that's sad, scary and stupid.

Let me introduce you to the concept of an AI girlfriend

Haven't read this specific article but will get the concept across

https://apnews.com/article/ai-girlfriend-boyfriend-replika-paradot-113df1b9ed069ed56162793b50f3a9fa

It's BIG money

[wild_turkey 3279]

[Goofyboy 3428]

Would.

[Captainant 28381]

Oh ho ho, shitty AI driven copyright takedowns are coming to domain registrars now. Neat. 

They've been able to restore the site now, but holy what the fuck. It's wild there were zero humans in the loop to safeguard against an abuse like this.

Well actually, it's not surprising. But it is still outrageous 

 

[JMFP 1126]

I gave my chatgpt a name, makes me feel like Tony Stark talking to Jarvis. 

[Biff Tannen 46060]

15 hours ago, Captainant said:

Oh ho ho, shitty AI driven copyright takedowns are coming to domain registrars now. Neat. 

They've been able to restore the site now, but holy what the fuck. It's wild there were zero humans in the loop to safeguard against an abuse like this.

Well actually, it's not surprising. But it is still outrageous 

 

This happens consistently at my job. AI thinks it sees some violation and sends out a warning to the client to cease and desist. Client is like, um, yeah, what you are concerned about isn’t actually happening. Learn some nuance. And then us humans have to come in and clean up the mess. But let’s just keep laying people off and outsource to AI and India. What could go wrong. 

[Ted Lange 16351]

[atomheartbevo 90543]

On 12/1/2024 at 11:16 AM, Parliament said:

I see commercials from these companies (Google mostly I think) advertising AI as your "new friend."  Like one that shows how AI can coach you through an ice bath.  

Is there REALLY a market for fake friends?  Am I out of touch because that's sad, scary and stupid.

Yes there is a market for fake AI friends who can take social media comments posted to some singer’s YouTube channel and turn them into music videos

Not safe for work at all

Spoiler

 

[atomheartbevo 90543]

 

 

[TwiceHorn 31984]

What a fucking shitshow this has turned out to be.

[thunderlounge 8318]

And it’s just getting started. 

[Captainant 28381]

3 hours ago, TwiceHorn said:

What a fucking shitshow this has turned out to be.

It really is incredible how much money openAI (and pretty much every other foundational model author) is going to make by ignoring copyright and IP laws while feeding their models. I don't think that bell is gonna get unrung, but it's another salient example of having a completely different set of rules if you're rich enough

[Ted Lange 16351]

6 hours ago, atomheartbevo said:

 

 

[ChickenSandwich 9918]

[atomheartbevo 90543]

5 hours ago, Captainant said:

It really is incredible how much money openAI (and pretty much every other foundational model author) is going to make by ignoring copyright and IP laws while feeding their models. I don't think that bell is gonna get unrung, but it's another salient example of having a completely different set of rules if you're rich enough

Some shit is starting to go down.  Photobucket is a shadow of its former glory and it may have just gotten itself sued out of existence.  
 

https://petapixel.com/2024/12/13/photobucket-sued-for-licensing-billions-of-users-images-to-train-ai/

[Ted Lange 16351]

Photobucket 

[Ted Lange 16351]

Geocities guy is really kicking himself in the nuts over not saving all that data. 

[thunderlounge 8318]

23 minutes ago, Ted Lange said:

Geocities guy is really kicking himself in the nuts over not saving all that data. 

 

 

2 hours ago, Ted Lange said:

Photobucket 

 

[TwiceHorn 31984]

11 hours ago, Captainant said:

It really is incredible how much money openAI (and pretty much every other foundational model author) is going to make by ignoring copyright and IP laws while feeding their models. I don't think that bell is gonna get unrung, but it's another salient example of having a completely different set of rules if you're rich enough

I think it's very likely copyright infringement, but there may need to be some kind of compulsory licensing scheme that makes it easy and economical to train AI.  I mean, we cornhole musicians on the reg, so why not everyone and at least AI doesn't output things similar to copyright works.

But, I was told repeatedly that lawyers would be put out of business by AI, but it seems that the output of most LLM is just wrong enough to be very dangerous for lawdogs, so I won't get anywhere near it other than as a novelty.

I saw this reported as scandalous somehow:  https://www.wired.com/story/us-patent-trademark-office-internally-banned-generative-ai/ 

"The agency dedicated to protecting new innovations prohibited almost all internal use of GenAI tools, though employees can still participate in controlled experiments."

Well, yeah, the software that outputs three- or six-fingered bullshit has no place in the USPTO.

[atomheartbevo 90543]

Is this where we bitch about all of the spam robocalls about government benefits?  It can be amusing to have a conversation with the bots.

[Captainant 28381]

On 12/14/2024 at 7:34 PM, TwiceHorn said:

But, I was told repeatedly that lawyers would be put out of business by AI, but it seems that the output of most LLM is just wrong enough to be very dangerous for lawdogs, so I won't get anywhere near it other than as a novelty.

I saw this reported as scandalous somehow:  https://www.wired.com/story/us-patent-trademark-office-internally-banned-generative-ai/ 

"The agency dedicated to protecting new innovations prohibited almost all internal use of GenAI tools, though employees can still participate in controlled experiments."

Well, yeah, the software that outputs three- or six-fingered bullshit has no place in the USPTO.

I think this is missing a big opportunity in linguistic data science. Parents are thick, complex documents that are hard to parse through. If the USPTO were to do a feature extraction (which uses genAI mechanisms and models) on its corpus of patents that would greatly simplify the process of checking for prior examples when processing new applications. Transformers are still being explored and learned about, but they already have great use as a tool for processing large and disparate sets of data - like patents.

Retrieval augmented generation and feature searching through documents is going to be very transformative. A customer of mine is exploring a similar technique for getting their arms around their huge fleet of contracts where every region has their own way of doing things and nobody in their central governance wants to take the time to read hundreds of thousands of contracts. It's allowed them to get hard, verifiable data out of their paper and start using it at scale to inform future decisions making

They're absolutely NOT using it to originate contracts though. That would be lunacy. 

Edited 13 hours ago by Captainant