Team-BHP - The Generative AI Thread (chatbot, text to image, text to video etc)
Team-BHP

Team-BHP (https://www.team-bhp.com/forum/)
-   Shifting gears (https://www.team-bhp.com/forum/shifting-gears/)
-   -   The Generative AI Thread (chatbot, text to image, text to video etc) (https://www.team-bhp.com/forum/shifting-gears/259867-generative-ai-thread-chatbot-text-image-text-video-etc-6.html)

Google announced Bard, a new experimental conversational Google AI service powered by LaMDA.

The Generative AI Thread (chatbot, text to image, text to video etc)-smartselect_20230207155815_twitter.jpg

The Generative AI Thread (chatbot, text to image, text to video etc)-smartselect_20230207155824_twitter.jpg

Introducing Bard

Quote:

It’s a really exciting time to be working on these technologies as we translate deep research and breakthroughs into products that truly help people. That’s the journey we’ve been on with large language models. Two years ago we unveiled next-generation language and conversation capabilities powered by our Language Model for Dialogue Applications (or LaMDA for short).

We’ve been working on an experimental conversational AI service, powered by LaMDA, that we’re calling Bard. And today, we’re taking another step forward by opening it up to trusted testers ahead of making it more widely available to the public in the coming weeks.

Bard seeks to combine the breadth of the world’s knowledge with the power, intelligence and creativity of our large language models. It draws on information from the web to provide fresh, high-quality responses. Bard can be an outlet for creativity, and a launchpad for curiosity, helping you to explain new discoveries from NASA’s James Webb Space Telescope to a 9-year-old, or learn more about the best strikers in football right now, and then get drills to build your skills.
Press Release

And then, they start it now for $20 a month (Bard effect?), which is ~1700 or a whopping ~20K a year (at the time of writing). Only question is - is it worthy?
The Generative AI Thread (chatbot, text to image, text to video etc)-p1.jpg

The Generative AI Thread (chatbot, text to image, text to video etc)-d1.jpg

In one browser (Brave) on my desktop which I used to sign in first, ChatGPT has 100% uptime. Anytime of the day it works well.

But on my laptop or my phone, ChatGPT asks me to check back later as the servers are busy, even though I'm using the same 'Google Account' everywhere for login.

Quote:

Originally Posted by SmartCat (Post 5491515)
In one browser (Brave) on my desktop which I used to sign in first, ChatGPT has 100% uptime. Anytime of the day it works well.

But on my laptop or my phone, ChatGPT asks me to check back later as the servers are busy, even though I'm using the same 'Google Account' everywhere for login.

Is it the same Brave browser you're using on your laptop and phone too? It wouldn't be surprising if MSFT eventually gives preferential treatment to Microsoft Edge users when accessing ChatGPT's free tier service while rate-limiting the access from other browsers.

Been following ChatGPT trends for a while now. Like any other AI bot, it is basically a trained algorithm, just that the amount of data fed to it is enormous, which is what puts it in the spotlight more than others so far. Likewise, there will be another bot in a few years that will surpass it, like ChatGPT did to IBM's Watson for example.

Whatever it answers is from the data that is used to train it. How it performs with unknown data (i.e. actual learning) is to be seen in the future. That is the real use case for any ML algorithm. Even then, there will be backtracking corrections like in any Neural Network which unfortunately involves validation by a human. This is the fundamental difference between an AI and a human - AI will have to be validated by a human.

As a result, ChatGPT is and will be biased, just because human beings are biased. True to that, this video pretty much confirms the suspicion:

https://www.youtube.com/watch?v=_Klkr6PtYzI

If we do not address this bias soon, ChatGPT will be useless in certain topics like Wikipedia is useless today in the context of politics, religion, and current affairs topics. It will be good in only those areas where it is easy to catch the errors, like S.T.E.M. and Geography.

As for coding, software engineers have used a vast amount of automated tools for a long time. A compiler, test framework, build pipelines like Jenkins, or even autocomplete in an IDE are also in a way tools to assist coders. However, the way ChatGPT pulls out a C++ code from some repository seems to me to be very helpful, because it reduces the effort to search for solutions.

But creating SQL statements from comprehension or writing straightforward python code is not new. I remember in Oracle they have BI and Policy Automation, which used to take plain English as input and spit out SQL statements (DDL including Entity-Relationship models and DML i.e. "SELECT"). As for python, Amazon's DynamoDB Workbench has been generating code in Python, Java, and Javascript already for the queries. Other code generators have been in use in AWS too. Of course, there is scope of improvement and ChatGPT might be better right now, but the idea is not new.

Regarding the F.U.D. surrounding ChatGPT, the bot itself says (in one of the examples from above) that it is there to "assist". Anybody thinking it can "replace" might just be reading too much into mainstream media (when they promote something to have a meteoric rise in a short time, it is a red flag for bubble). If a task can be done by a robot, a human should not do it and move on to something smarter.

Bard gets an answer wrong, Google inadvertantly uses the same in its promotional ad; Alphabet loses $100 Billion in market value.

Link

ChatGPT has its first big kill!

Microsoft announces the new Bing with ChatGPT built in:

https://techcrunch.com/2023/02/07/mi...-built-in/amp/

However, only part of the features are built in and for the full experience you have to join a waitlist.

Probably amongst the best ROI anyone has got on an investment - $10 bn in ChatGPT reduces Google valuation by $100 bn. lol:

New York Times tech columnist writes:

"I'm switching my desktop computer's default search engine to Bing.And Google, my default source of information for my entire adult life, is going to have to fight to get me back.”


https://trib.al/ToJlmed?fbclid=IwAR3...8yrbXykhG3OuD0

ChatGPT can be used to make travel/road trip plans:

The Generative AI Thread (chatbot, text to image, text to video etc)-screenshot_3.png

The Generative AI Thread (chatbot, text to image, text to video etc)-screenshot_4.png

The Generative AI Thread (chatbot, text to image, text to video etc)-screenshot_5.png

And suggest list of clean restaurants in the vicinity:

The Generative AI Thread (chatbot, text to image, text to video etc)-screenshot_6.png

I think they need to quickly bring the data upto speed, this 2021 limitation wont work beyond a point for serious users. I have a feeling though they may have already corrected in the new Bing app which MS insists to make use of as default.

I asked ChatGPT a general question about a small Shopify based website that was created in 2022, May.

It told me exactly what it is about.

Is this Chat GPT off the grid? Not able to get some help from it for the past few days.

Shows this below message on the screen. Something changed?

Quote:

ChatGPT is at capacity right now. Get notified when we're back.

Quote:

Originally Posted by balenoed_ (Post 5496768)
Is this Chat GPT off the grid? Not able to get some help from it for the past few days.

Shows this below message on the screen. Something changed?

It’s available as part of bing now ( with some waitlist to join) in case you want to try out that way.

I know that chatbots aren’t sentient. But this is creepy to say the least. From NYT.

Extremely good read. And frightening.

After accepting that it is only a language neural program, I don’t know what to say about our society and internet in general. After all, it learnt everything from the data out there.

Is #MeToo tag applicable when an AI program makes unwanted advances? I can see American lawyers smiling in glee.


All times are GMT +5.5. The time now is 10:09.