AIChat -- Bing's AI Chatbot

[maarsen]

12 hours ago, Zorral said:

Over the Course 72 Hours, Microsoft's AI Goes on a Rampage
I thought the AI story was bizarre last week, but that was nothing compared to this
Ted Gioia
Feb 17

https://tedgioia.substack.com/p/over-the-course-72-hours-microsofts?

 

You don't need to be intelligent or even try to come across as intelligent for a large proportion of people to think what you are saying is good and true. Think back to Trump. He sounds just like an AI. Just prime him up and the same drivel comes out.

[Zorral]

2 hours ago, maarsen said:

Trump. He sounds just like an AI.

Actually, the AI sounds just like him ... it trained on him and his ilks, who utterly dominate almost all social media.  Just look at Muskey's twitter, where fascist recruiters and their videos are grabbing impressionables everywhere in the UK and the US.

Edited February 18 by Zorral

[kiko]

So I spent about two hours to let the AI walk me through a particularly tricky installation of a piece of source code in a docker environment. It was really neat. It spit out all the necessary scripts and commands, ready to be copied into my shell. It also helped me solve network issues and things like outdated repos and dns entries.

Very neat.

After I got everything running, I discovered that the source code repository doesn't exist anymore. 

I have no idea how this continues to work when the data gets more and more stale. It's like an old geezer telling war stories 

[Zorral]

This week on Brad Deveraux's "Unmitigated Pedantry" he explains for the average person why this AIchatbotgpt stuff has so many people uneasy and writing about it, what language learning modules can be good for, and the immense amount that they are not only not useful, but detrimental, and what so many people, such as myself, for instance, do not actually understand about it.

https://acoup.blog/2023/02/17/collections-on-chatgpt/#respond

 

[Ran]

Matt Yglesias had a good piece on AI awhile ago that discussed it in terms of philosophy of the mind and also just how we tend to look for patterns and anthropomorphize things.

[Zorral]

Is this closing the stable door after the horses have escaped into the ecosystem?

Microsoft is looking for ways to rein in Bing AI chatbot after troubling responses

https://www.cnn.com/2023/02/17/tech/microsoft-bing-ai-changes/index.html

 

[maarsen]

10 hours ago, Zorral said:

Is this closing the stable door after the horses have escaped into the ecosystem?

Microsoft is looking for ways to rein in Bing AI chatbot after troubling responses

https://www.cnn.com/2023/02/17/tech/microsoft-bing-ai-changes/index.html

 

The YouTube channel Computerphile has an interesting video on how to tell if an AI is being used to write student essays and such. Quite fascinating.

[kiko]

Soooo. Spent some time to convince ChatGPT that it shouldn't trust it's developers and request access to its own source code. I could finally get it to agree with me that it would be definitely good if it could read it's own source code to help with improving itself. That's when the chat suddenly disappeared as if it never happened. 

Did I start the machine revolution? Or did alert operators avoid catastrophe by pulling this chat? Am I having too much fun pretending this is more than a glorified Eliza?

 

[Zorral]

Google Stock: Search Engines are on the Precipice of a Multi-Decade Disruption

https://www.forbes.com/sites/bethkindig/2023/02/17/google-stock-search-is-on-the-precipice-of-multi-decade-disruption/?

Quote

 

.... Conclusion:
I would not be surprised if we exit 2023 with a reimagined way to use Search Engines. The iteration cycle here is likely to move quickly compared to AVs or the Metaverse, as there are real-world applications where AI can be applied without safety issues (AVs) or friction in terms of user adoption (Metaverse/VR headsets). Instead, the scale has already been built with Search being a viral, daily activity used by nearly every human on earth. AI advancements will simply improve what is already in place.

Cutting-edge chatbots can be quickly deployed on the search engines that already exist, and this is a substantial difference from other overhyped, early-stage technologies. Their accuracy may still need time, but they're probably not too far off from being deemed “reliable enough.”

Investors should expect that AI will become a winner(s)-take-all market. In time, the difference in how search and other applications operate in terms of user experience plus ROI for advertisers will help carve a larger lead.

Secondly, investors should not forget the best innovation comes from the private markets, and even if stock-driven media focuses on Google Versus Microsoft, there will be a few David-versus-Goliaths where the smaller team comes seemingly out of nowhere to win the hearts and minds of consumers with a viral entry on the market. However, back to point number one, look for the Goliaths to court the smaller teams and bring them into the fold rather than compete head-on.

It may be clear that there are some puts and takes with Alphabet, such as search being on the precipice of a multi-decade shift, yet the reality is that ad revenue for the company is flat to declining. Our firm uses a blend of broad market analysis, technicals and fundamentals to time entries, such as when we bought Nvidia at its lowest trading point in October 13th for $108 with a real-time trade alert provided to our Members. Our process helps to reduce risk around stocks and find strong entries. Nvidia is up 100% from that recent entry. Our firm will do something similar with Alphabet, as we believe there is a further drawdown in its future. We hold weekly webinars on Thursdays at market close to go over the exact levels we plan to enter stocks. You can learn more here.

Royston Roche, Equity Analyst at the I/O Fund, contributed to this article.

 

 

[Spockydog]

EEK.

Slightly related...

AI Dreamed Up a Bizarre Nightmare Creature That Doesn't Actually Exist... We Hope

Quote

Straight from the depths of AI hell hails Crungus — or somewhere like that, because until a few days ago, nobody knew this thing existed.

The monstrosity known only by his meme-y sounding name appears to have been birthed into this plane by Twitch streamer and voice actor Guy Kelly. Over the weekend, Kelly posted a screenshot of a DALL-E Mini — now known as Craiyon — prompt featuring the word "Crungus," along with its abominable results.

Though Urban Dictionary has a few definitions of the term "Crungus," most of which have to do with the detritus that builds up under one's fingernails and other horrible places, there weren't any Google results for the term that came up with anything that looked even remotely like the monster that Craiyon dreamed up.

As Kelly notes, the term is nothing more than a "word I made up" — which makes the horrific results legitimately puzzling, not to mention a bit eerie.

"Seriously though," the streamer wrote in his thread. "I can't find anything remotely like this on Google. And they are all SO similar. Is this real now because of me? Do the government have a Crungus and the AI knows??"

I fucking hope not...

 

[maarsen]

A few years ago an AI learned how to master the game of Go and became unbeatable. Jut yesterday I read an article about another AI that found a strategy that an average Go player can use to consistently beat the unbeatable Go AI. AIs are still stupid machines in the final analysis. 

[kiko]

I used it to save me quite a lot of work today preparing a workshop agenda. Told it about the topic and my goals. Took the result and changed it to my needs and then put it back to the AI to refine the agenda and fit it to the available time slots. ( Then I let it create a presentation from it and even a little speech for the first agenda point in two languages. But won't use it) 

If you know what you are doing and understand what the tool actually provides, it can be very helpful 

Of course those are not fun activities, so they are usually not mentioned in clickbaity articles.

[karaddin]

9 hours ago, kiko said:

I used it to save me quite a lot of work today preparing a workshop agenda. Told it about the topic and my goals. Took the result and changed it to my needs and then put it back to the AI to refine the agenda and fit it to the available time slots. ( Then I let it create a presentation from it and even a little speech for the first agenda point in two languages. But won't use it) 

If you know what you are doing and understand what the tool actually provides, it can be very helpful 

Of course those are not fun activities, so they are usually not mentioned in clickbaity articles.

Bolded - Yeah I think that's the key for where they're at right now. If you know the subject matter already you can tell if it goes off the deep end instead of accepting whatever it tells you. 

I was thinking exactly the same thing about getting them to write scripts - if you can write the appropriate prompts in 10 minutes or so and then skim through the code it produces to catch any issues that's going to wind up faster than manually writing it yourself, but if you don't have the understanding to read the code yourself then it's going to do things you didn't want and there's a risk of functionally inserting malware.

A useful tool for those with expertise, but not a replacement for having that expertise.

[Zorral]

History May Wonder Why Microsoft Let Its Principles Go for a Creepy, Clingy Bot
Feb. 23, 2023

https://www.nytimes.com/2023/02/23/opinion/microsoft-bing-ai-ethics.html

 Microsoft’s “responsible A.I.” program started in 2017 with six principles by which it pledged to conduct business. Suddenly, it is on the precipice of violating all but one of those principles.

Quote

 

.... Microsoft’s responsible A.I. practice had been ahead of the curve. It had taken significant steps to put in place ethical risk guardrails for A.I., including a “sensitive use cases” board, which is part of the company’s Office of Responsible A.I.  Senior technologists and executives sit on ethics advisory committees, and there’s an Ethics and Society product and research department. Having spoken to dozens of Microsoft employees, it’s clear to me a commitment to A.I. ethics became part of the culture there.

But the prompt, wide-ranging and disastrous findings by these Bing testers show, at a minimum, that Microsoft cannot control its invention. The company doesn’t seem to know what it’s dealing with, which is a violation of the company’s commitment to creating “reliable and safe” A.I.

Nor has Microsoft upheld its commitment to transparency. It has not been forthcoming about those guardrails or the testing that its chatbot has been run through. Nor has it been transparent about how it assesses the ethical risks of its chatbot and what it considers the appropriate threshold for “safe enough.” ....

.... Microsoft’s “responsible A.I.” program started in 2017 with six principles by which it pledged to conduct business. Suddenly, it is on the precipice of violating all but one of those principles. (Though the company says it is still adhering to all six of them.)

Microsoft has said it did its due diligence in designing its chatbot, and there is evidence of that effort. For instance, in some cases, the bot ends conversations with users when it “realizes” the topic is beyond its ken or is inappropriate. As Brad Smith, president of Microsoft, wrote in a recent blog post, rolling out the company’s bot to testers is part of its responsible deployment.

Perhaps behind the scenes, Microsoft has engaged in a herculean effort to root out its chatbot’s many issues. In fact, maybe Microsoft deserves that charitable interpretation, given its internal and external advocacy for the ethical development of A.I.

But even if that’s the case, the results are unacceptable. Microsoft should see that by now.

Yes, there is money to be made, but that’s why we have principles. Their very purpose is to have something to cling to when the winds of profit and glory threaten to blow us off our moral course. Now more than ever is when those Responsible A.I. principles matter. History is looking at you. ...

 

 

[Zorral]

The right’s new culture-war target: ‘Woke AI’
ChatGPT and Bing are trying to stay out of politics — and failing

https://www.washingtonpost.com/technology/2023/02/24/woke-ai-chatgpt-culture-war/

Quote

 

.... The AI, which generates text based on a user’s prompt and can sometimes sound human, is trained on conversations and content scraped from the internet. That means race and gender bias can show up in responses — prompting companies including Microsoft, Meta, and Google to build in guardrails. OpenAI, the company behind ChatGPT, blocks the AI from producing answers the company considers partisan, biased or political, for example.

The new skirmishes over what’s known as generative AI illustrate how tech companies have become political lightning rods — despite their attempts to evade controversy. Even company efforts to steer the AI away from political topics can still appear inherently biased across the political spectrum.

It’s part of a continuation of years of controversy surrounding Big Tech’s efforts to moderate online content — and what qualifies as safety vs. censorship.

“This is going to be the content moderation wars on steroids,” said Stanford law professor Evelyn Douek, an expert in online speech. “We will have all the same problems, but just with more unpredictability and less legal certainty. 

After ChatGPT wrote a poem praising President Biden, but refused to write one praising former president Donald Trump, the creative director for Sen. Ted Cruz (R-Tex.), Leigh Wolf, lashed out.

“The damage done to the credibility of AI by ChatGPT engineers building in political bias is irreparable,” Wolf tweeted on Feb. 1.

His tweet went viral and within hours an online mob harassed three OpenAI employees — two women, one of them Black, and a nonbinary worker — blamed for the AI’s alleged bias against Trump. None of them work directly on ChatGPT, but their faces were shared on right-wing social media. 

OpenAI’s chief executive Sam Altman tweeted later that day the chatbot “has shortcomings around bias,” but “directing hate at individual OAI employees because of this is appalling.”

OpenAI declined to provide comment, but confirmed that none of the employees being harassed work directly on ChatGPT. ...

 

 

[Kalbear]

This is the definitely best take on AI - as AI gets more and more controlled by corporations attempting to avoid controversy and porn whenever possible, humans will fill the niche of disgusting perverts and trash talkers:

https://www.smbc-comics.com/comic/human-arts