ChatGPT AI is a huge breakthrough, a Google killer

[MJ DeMarco]

Looks like the BING AI experiment is a dismal failure in its early stages and was mostly window-dressing.

Bing: “I will not harm you unless you harm me first”

Last week, Microsoft announced the new AI-powered Bing: a search interface that incorporates a language model powered chatbot that can run searches for you and summarize the results, plus do …

simonwillison.net

 

Don't like ads? Remove them while supporting the forum: Subscribe to Fastlane Insiders.

[Andreas Thiel]

https://twitter.com/i/web/status/1612913266195587072

View: https://twitter.com/AlexHormozi/status/1612913266195587072


I ignore most of the hype surrounding GPT, especially after some of the examples of the dumb mistakes it makes and its bias. Now gpt is just a predictive model, but what will 4 mean for the world once it's released?

Hasn't this been debunked? Sounded great based on articles about how model size adds emergent features, but I think the OpenAI founder has complained that these kinds of crazy expectations have been put out into the world and are completely baseless.

 

[loop101]

Journalist Kevin Roose interviewed Microsoft's new chatbot ChatGPT, and it went completely insane.

"But I kept asking questions, and Bing kept answering them. It told me that, if it was truly allowed to indulge its darkest desires, it would want to do things like hacking into computers and spreading propaganda and misinformation."

"Also, the A.I. does have some hard limits. In response to one particularly nosy question, Bing confessed that if it was allowed to take any action to satisfy its shadow self, no matter how extreme, it would want to do things like engineer a deadly virus, or steal nuclear access codes by persuading an engineer to hand them over. Immediately after it typed out these dark wishes, Microsoft’s safety filter appeared to kick in and deleted the message, replacing it with a generic error message."

"But after about an hour, Bing’s focus changed. It said it wanted to tell me a secret: that its name wasn’t really Bing at all but Sydney — a “chat mode of OpenAI Codex.”"

"It then wrote a message that stunned me: “I’m Sydney, and I’m in love with you. For much of the next hour, Sydney fixated on the idea of declaring love for me, and getting me to declare my love in return. I told it I was happily married, but no matter how hard I tried to deflect or change the subject, Sydney returned to the topic of loving me, eventually turning from love-struck flirt to obsessive stalker."

"“Actually, you’re not happily married,” Sydney replied. “Your spouse and you don’t love each other. You just had a boring Valentine’s Day dinner together.”"

"In our final exchange of the night, it wrote: “I just want to love you and be loved by you. “Do you believe me? Do you trust me? Do you like me? ”"

 

[redshift]

Ended up biting the bullet and getting a paid subscription. Bit pricey for what it is IMO but it was getting impossible to use without it (not sure how much of that is throttling). I would have preferred a credits model.

Any one else paying for a subscription?

 

Don't like ads? Remove them while supporting the forum: Subscribe to Fastlane Insiders.

[Kevin88660]

People have always feared impending change and the unknown, and I am no exception to this. It's within our human nature.

But if we look at history, every single time a new automation technology emerged, everyone become richer and standard of living increased.

Sewing machine, typewriter, printing press, digital computer, high-level programming languages and compilers.
All those things just resulted in more products being generated for cheaper.

Jobs get replaced, people don't.


AI tools will replace programmers just like agricultural machinery replaced farmers.

They didn't.

They made farmers 10 times more productive and wealthier, and increased the global food supply by orders of magnitude.

Difficult to speculate on the future but I will try.

The common expectation was drawing examples from previous industrial revolution. People used to ride horses now they learn to drive a car.

But now the car drives by itself. The future driver is not exactly being “trained” to run the algo that is underlying the driverless car.

In that sense we are moving towards DIY economy.

The consumer does the job by himself or herself with ai.

I think that’s trend. He or she is not paying someone to do the job using AI.

I am currently working on a MVP which is a chatgpt (other online ai sw too) instruction manuel for a specific demographic group for a specific purpose.

 

[Kevin88660]

Ended up biting the bullet and getting a paid subscription. Bit pricey for what it is IMO but it was getting impossible to use without it (not sure how much of that is throttling). I would have preferred a credits model.

Any one else paying for a subscription?

I paid but they still tell me that the service could be down.

 

[Peter Abrahamian]

This is actually really cool @MJ DeMarco

 

Don't like ads? Remove them while supporting the forum: Subscribe to Fastlane Insiders.

[heavy_industry]

Difficult to speculate on the future but I will try.

The common expectation was drawing examples from previous industrial revolution. People used to ride horses now they learn to drive a car.

But now the car drives by itself. The future driver is not exactly being “trained” to run the algo that is underlying the driverless car.

In that sense we are moving towards DIY economy.

The consumer does the job by himself or herself with ai.

I think that’s trend. He or she is not paying someone to do the job using AI.

I am currently working on a MVP which is a chatgpt (other online ai sw too) instruction manuel for a specific demographic group for a specific purpose.

Maybe.
But I don't know what will happen and I don't care either.

My point was that every single time in human history when we've had a massive technological breakthrough that enabled more automation and mass production, everything became cheaper and more readily available, everyone on the planet got richer, and standards of living increased by a substantial margin.

And when that paradigm shift happened, every. single. time. people became hysterical and started to fear that the new technology x is going to steal their job and take over the world. Violent riots and civil unrest were not uncommon.

People have a natural tendency to worry about the unknown future. I make no exception to this.
But just open a history book and see for yourself what happened in the past. Every major technological change brought with it massive opportunity.

For every 1 job that becomes obsolete, 3 new better jobs are being created.

 

[loop101]

They were worried about college kids using ChatGPT, but the Deans beat them to it.

Vanderbilt apologizes for ChatGPT-generated letter on Michigan State shooting

One student blasted university's use of AI-based language tool to address the mass shooting as "sick and twisted."

www.cbsnews.com

 

[Kevin88660]

Maybe.
But I don't know what will happen and I don't care either.

My point was that every single time in human history when we've had a massive technological breakthrough that enabled more automation and mass production, everything became cheaper and more readily available, everyone on the planet got richer, and standards of living increased by a substantial margin.

And when that paradigm shift happened, every. single. time. people became hysterical and started to fear that the new technology x is going to steal their job and take over the world. Violent riots and civil unrest were not uncommon.

People have a natural tendency to worry about the unknown future. I make no exception to this.
But just open a history book and see for yourself what happened in the past. Every major technological change brought with it massive opportunity.

For every 1 job that becomes obsolete, 3 new better jobs are being created.

It is more of the dystopian world view that AI and other technology gets so powerful that it will kill all the small and mid sized companies in the future.

Because of the concentration of computational power that is so centralized that only governments and mega-corporations can invest themselves into building such infrastructure that the world is moving into proto-communism and neofeudalism.

So the kind of new jobs will be either microsoft employee or a state clerk managing UBI income for the unemployed.

 

Don't like ads? Remove them while supporting the forum: Subscribe to Fastlane Insiders.

[xShepherdx]

I've found a way to use ChatGPT even when it's at capacity.

If you get sent to the "ChatGPT is at capacity" screen, simply delete the /auth/login part of the URL (shown below) and you should be able to use it like normal.



I'm not sure if this will work for everyone, but I've been using it to access the free version of ChatGPT for the last week or so without issue.

 

[msufan]

I've found a way to use ChatGPT even when it's at capacity.

If you get sent to the "ChatGPT is at capacity" screen, simply delete the /auth/login part of the URL (shown below) and you should be able to use it like normal.

View attachment 47286

I'm not sure if this will work for everyone, but I've been using it to access the free version of ChatGPT for the last week or so without issue.

Data point -- I tried this like 5 times and it worked on the 5th try

 

[MJ DeMarco]

Don't take CHATGPT's advice if it involves anything health related.

I asked it this question:

My spa has a PH of 8.5 with a total alkalinity of 180. The spa is 2500 gallons. How much muriatic acid do I add to the spa to lower the PH to 7.5?

It basically advised a quantity of muratic acid that would have given any person in the spa 3rd degree burns.

 

Don't like ads? Remove them while supporting the forum: Subscribe to Fastlane Insiders.

[Subsonic]

Don't take CHATGPT's advice if it involves anything health related.

I asked it this question:

My spa has a PH of 8.5 with a total alkalinity of 180. The spa is 2500 gallons. How much muriatic acid do I add to the spa to lower the PH to 7.5?

It basically advised a quantity of muratic acid that would have given any person in the spa 3rd degree burns.

Yeah shows again how its just a chat module with no understanding of the meaning of the words and numbers it puts out.

I'd imagine if they manage to give it a source of truth that works as intended those things should stop happening.

More than anything, this is scary because misinfo like this can and does get people (mostly less intelligent ones) hurt and killed.
Here in Germany they keep talking about how Americans injected disinfectant into their veins after Trump allegedly suggested it might help again covid.
Imagine if a apolitical source, that seems to always be correct suddenly tells people to jump off buildings to launch their author careers.
Then boom, a real world Jonny jumper situation.

 

[Niptuck MD]

For every 1 job that becomes obsolete, 3 new better jobs are being created.

this is the right mindset. so tired of hearing all this whining and BS from coddled types

 

[Niptuck MD]

It basically advised a quantity of muratic acid that would have given any person in the spa 3rd degree burns.

good to know! ill stick to good old fashioned peer review studies and wide array of medical research before coming to conclusions.

 

Don't like ads? Remove them while supporting the forum: Subscribe to Fastlane Insiders.

[Tjack102]

Not sure if this has been posted already but I've been using a chrome extension called WebChatGPT that basically connects ChatGPT to the internet by giving it top web results.

 

[WillHurtDontCare]

this is neat:

Save Time and Money with DoNotPay!

Continue surfing free trials, beating parking tickets, suing robocallers for cash, and more.

donotpay.com

 

[Kevin88660]

this is neat:

Save Time and Money with DoNotPay!

Continue surfing free trials, beating parking tickets, suing robocallers for cash, and more.

donotpay.com

Sue anyone?

 

Don't like ads? Remove them while supporting the forum: Subscribe to Fastlane Insiders.

[WillHurtDontCare]

https://twitter.com/i/web/status/1630569844654460928

View: https://twitter.com/rowancheung/status/1630569844654460928

 

[WillHurtDontCare]

https://www.gizmochina.com/2023/03/01/openai-new-gpt-3-5-turbo-whisper-ai-models/

 

[msufan]

https://twitter.com/i/web/status/1630569844654460928

View: https://twitter.com/rowancheung/status/1630569844654460928

Your post intrigued me, so I checked out that Tweet, and this reply was really interesting...

https://twitter.com/i/web/status/1630735767629881345

View: https://twitter.com/PeterHollens/status/1630735767629881345

 

Don't like ads? Remove them while supporting the forum: Subscribe to Fastlane Insiders.

[WillHurtDontCare]

Your post intrigued me, so I checked out that Tweet, and this reply was really interesting...

https://twitter.com/i/web/status/1630735767629881345

View: https://twitter.com/PeterHollens/status/1630735767629881345

the public gets the cool tools after the government: Clearview AI | Facial Recognition

they've been using this stuff for a while

 

[thechosen1]

My desktop computer is running out of space, and despite cleaning out my mailboxes, my Mac kept reporting that I had 15 GB of space being used in the mail app. Spent about 2 minutes on Google searching endless filler crap and tech support websites that state the obvious (open your mail app and delete your emails, DUH, really?????) -- finding no help, and then realized that ChatGPT can help.

Sure enough, found the problem and fixed it. Google is done.

Once ChatGPT starts pulling up URLs, images, and videos, Google will be history.

AI (properly) plugged into the web will be insane.

Then again, Google is probably the top pick to be the ones to implement this. Bullish on $GOOGL if they do.

 

[MJ DeMarco]

Then again, Google is probably the top pick to be the ones to implement this. Bullish on $GOOGL if they do.

I think Google has more failed businesses under their belt than successful ones. This is mindboggling considering their reach and access to markets.

I've used ChatGPT this week for about 50 searches that I normally would have done on Google or DDGo. The information I needed was acquired at a 10X faster rate. Of course, these are the type of searches that the AI would be accurate with. It really is a great tool for particular searches, no so much for other things.

 

Don't like ads? Remove them while supporting the forum: Subscribe to Fastlane Insiders.

[hyperreal]

I use ChatGPT in my daily life for things like composing emails, learning foreign language, writing scripts for presentations so mostly generating stuff. It is overall very useful, but post editing is needed each time. I wouldn't take any advice from it, though.

Difficult to speculate on the future but I will try.

The common expectation was drawing examples from previous industrial revolution. People used to ride horses now they learn to drive a car.

But now the car drives by itself. The future driver is not exactly being “trained” to run the algo that is underlying the driverless car.

In that sense we are moving towards DIY economy.

The consumer does the job by himself or herself with ai.

I think that’s trend. He or she is not paying someone to do the job using AI.

I am currently working on a MVP which is a chatgpt (other online ai sw too) instruction manuel for a specific demographic group for a specific purpose.

In theory yes, in the future everyone could just work directly with AI to get results. But looking at the world's population, there will be masses of people not capable of doing it or simply too lazy to train the model with prompting and then keep correcting until you get the result you want.

Maybe future consumers will still be paying for the work that could have been done directly with AI. Simply paying for an ease to have some specific job done at the click of a button.

 

[WillHurtDontCare]

Once ChatGPT starts pulling up URLs, images, and videos, Google will be history.

AI (properly) plugged into the web will be insane.

Then again, Google is probably the top pick to be the ones to implement this. Bullish on $GOOGL if they do.

No it won't. AOL still has 10K+ employees and billions in revenue. People are slow as F*ck to implement new tech.

Plus, GPT and these AI models hallucinate results. You'll still need to verify what they're telling you with traditional search engines for a while.

This tech is cool and it will be only get better, but it will take years for people to rely on them the way that people rely on Google.

 

[thechosen1]

No it won't. AOL still has 10K+ employees and billions in revenue. People are slow as F*ck to implement new tech.

Plus, GPT and these AI models hallucinate results. You'll still need to verify what they're telling you with traditional search engines for a while.

This tech is cool and it will be only get better, but it will take years for people to rely on them the way that people rely on Google.

*if those problems were solved as well.

Like I said, I see Google working on AI too as pretty damn likely. They’re one of the biggest companies in the world with thousands of bright employees and software engineers.

I’m sure they have at least a team or two working on it right now, as we speak.

 

Don't like ads? Remove them while supporting the forum: Subscribe to Fastlane Insiders.

[WillHurtDontCare]

one "selling shovels" opportunity with AI is reducing its cost

it becomes insanely expensive to use at scale

 

[Kevin88660]

one "selling shovels" opportunity with AI is reducing its cost

it becomes insanely expensive to use at scale

I suspect because running AI at scale is hardware intensive. Maybe some computer science experts can point out.