Always strive to learn something useful. --Sophocles

*sigh*

Care to guess how it happened? The suggestions included "Tidewater Dreams" by Isabel Allende and "The Last Algorithm" by Andy Weir". The independent who put the list together used an AI and didn't check what it generated.

The Sun-Times went through some massive lay-offs recently as its finances are in not very good shape, and lost 20% of its readership. I'm sure this little reading list snafu will encourage people to reup their subscriptions. Or not.

https://arstechnica.com/ai/2025/05/chicago-sun-times-prints-summer-reading-list-full-of-fake-books/

This is pretty amusing, funny, and ironic, for certain values of amusing, funny, and ironic.

The Chinese have released an open source LLM chatbot, it's called DeepSeek R1. It's available on GitHub, you can download it and play with it, tear apart the code, tweak it, etc. Completely free. It was built in TWO MONTHS for $6,000,000. And on lower grade GPUs because of export restrictions placed on the Chinese - can't have them getting top of the line GPUs now, can we? And they didn't scrape the internet, stealing trademarked and copyrighted information without permission to train it.

And it's the top downloaded app on the Apple app store.

The belief was that with denying the Chinese the top-end GPUs needed to do the crunching to construct large language models, that they had no hope of catching up with Team USA when it came to building AIs. Well, it seems that Team USA forgot the phrase 'Work smarter, not harder'. The Chinese applied a lot of smarts to work around the restrictions that were placed upon them and produced a much better product.

And stocks tanked because it's good software. From the article: "GPU maker NVIDIA fell 11%, Oracle dropped 8%, and Palantir was down 5% ... Stocks are adjusting to the revelation that China can build AI faster, cheaper, and just as good as America."

The comments in the article are amusing.

Nice slice of humble pie served up there.

https://gizmodo.com/chinese-ai-deepseek-deep-sixes-openai-on-the-app-store-stocks-tank-2000555171

This is interesting.

The company who, up until 2012, published the book-form of the Encyclopedia Britannica, is now turning that huge trove of facts into an LLM engine with the goal of selling it as a service to the education market.

While this might seem as a bit of a snoozer, there's one very interesting aspect to this: AI hallucinations.

Most LLM models have hallucination problems, seemingly stemming from their snarfing up their training data from hoovering up the internet with all of its crappy and contradictory information. This is where Britannica shines: they paid a literal fortune over two centuries collecting vetted materials from recognized scholars using quality editors to compile it into a trusted source. Thus, the quality of their training model will be very, very high.

The question will be if their code that ingests this training model will still hallucinate. And we'll only see that with testing when it goes public and really gets pummeled. But I do like the idea: starting with a very high quality training set, I think it shows promise.

Though we still have the problem of AI systems consuming stupid godawful amounts of energy.

Britannica's encyclopedia is still available online, just not in a print edition.

https://gizmodo.com/encyclopedia-britannica-is-now-an-ai-company-2000542600

First off, it has to be pointed out this is a specialized AI model designed for programmers, not a generalized model like ChatGPT et al.

IBM trained it specifically on open source libraries to which they explicitly had permission, basically bending over backwards to avoid any possible legal issues. And they now have a working model that they've released to the public! Granite was trained on 116 different programming languages and has from 3 to 34 billion tokens, presumably per language. I wonder if you can ask it to list all the languages it's trained in, I'll bet there's some pretty esoteric ones in there! I'd love it if it had MUMPS! (I once found a book on MUMPS programming at the Phoenix Public Library, I imagine it's been weeded by now)

Anyway, interesting article. It describes how it was trained, etc., but one of the more interesting bits was saying that in the rather short time since ChatGPT et al have appeared and everyone started creating their own LLMs, the cost for training up an LLM has dropped from millions of dollars to thousands! That's a pretty impressive scale drop.

https://www.zdnet.com/article/ibm-open-sources-its-granite-ai-models-and-they-mean-business/

https://www.zdnet.com/article/ibm-open-sources-its-granite-ai-models-and-they-mean-business/

So what?, many of you say.

He had a near fatal stroke eleven years ago and almost completely lost the ability to speak and sing! In that time, he's done some acoustic playing and was inducted into the CW Hall of Fame. He's not an invalid, no idea how much the stroke affected him otherwise.

Details have not been fully released, but through using AI technology, they've recreated his voice and he has a new song out.

Now obviously this brings up a host of issues. I am absolutely okay with this: Randy seems to be fully competent of mind and body, just not able to speak or sing. He has control of his music, and participated in this project to recreate his voice and get this song out. I think this is quite awesome! Not that I'm a CW fan or really care one way or another, but it's great for an artist to be able to express themselves creatively after nearly dying from a stroke and being deprived of their greatest instrument.

But you can see where this can be abused. As much as I'd love to hear new Freddie Mercury, or David Bowie, they're long gone and cannot participate in the process. No input from them. Theoretically their estate or IP managers could, and I would think that's wrong. I have the same problem with cinema recreations of James Dean or Humphrey Bogart. Digital de-aging of people, such as the ABBA Voyager holographic performance? Very interesting. And all four of them are alive to consent to it. Digital recreations of a performing Tupac? Very problematic for me.

One thing that we don't know about the Randy Travis thing which may come out eventually is who performed the vocals for the recording, which is being supposed that Randy's generated voice was then layed over.

https://www.rollingstone.com/music/music-country/randy-travis-releases-ai-song-1235014871/

And what's even harderer? Using an AI coding assistant to write secure programs.

Many, MANY times that I've written about computer insecurity issues I've said explicitly that computer security is HARD. And here we have a prime example.

It turns out that using an AI to help you write a program produces LESS secure programs! But that's not the worse part: the program is more likely to believe that they are writing MORE SECURE CODE!

This is very bad. I've used AI for hints in writing code, looking for little obscure code references that I'm not familiar with. Quite useful. I haven't used it to write entire programs for me, I'm not sure that I could. However, there are people out there paying for subscriptions to ChatGPT 4 and other engines using them heavily, and that is worrisome.

https://arxiv.org/html/2211.03622v3

https://www.schneier.com/blog/archives/2024/01/code-written-with-ai-assistants-is-less-secure.html

Dropbox has an OPT-OUT 'feature' that they've just implemented where any files that you upload to Dropbox for storage or sharing are automatically shared with OpenAI. They claim it is NOT for AI training. YAY?! But then why are they sharing it with the best-known AI company? Doesn't make sense.

Supposedly Dropbox is developing an AI-assisted cross-platform search feature that will let you search for stuff across Dropbox, Outlook email, etc. And this is in preparation for it. However, it is turned on by default. To turn it off, you have to sign on to Dropbox through a web browser and dig into your settings.

The Ars article has instructions and a link on how to deactivate it. Allegedly your data will be discarded before 30 days.

Myself, I stopped using Dropbox a couple of years ago when they capped the number of devices that could link to an account to three. While I liked the ease with which I could share links to a file, I didn't find it worth paying for when there were free options that also had file sharing.

An amusing thing about comments in the Ars article: the number of people closing their Dropbox accounts over this!

https://arstechnica.com/information-technology/2023/12/dropbox-spooks-users-by-sending-data-to-openai-for-ai-search-features/

As of Wednesday, Altman was back at OpenAI and most of the board that fired him were gone.

Satya Nadella, CEO of Microsoft, who allegedly had hired Altman and others who had fled OpenAI to preside over and to staff an AI special projects division, was fully supportive of Altman and his Greyhound Bus of followers to return to their old lodgings. MS HR heaved a collective sigh of relief that none of the paperwork was ever actually sent down.

Three of the four board members who fired Altman last week are gone. From the article, and emphasis mine: "The one OpenAI board member who is staying is Quora CEO Adam D'Angelo, who was reportedly involved in the discussions that led to Altman's return. The three who are leaving the board are OpenAI Chief Scientist Ilya Sutskever, entrepreneur Tasha McCauley, and Helen Toner of the Georgetown Center for Security and Emerging Technology."

You will remember Ilya Sutskever from previous posts as being initially vocal about the need for Altman to go, then vocal about the need for him to return and being number twelve, IIRC, on the letter ultimately signed by 95% of OpenAI employees threatening to quit if Altman did not return and the board resign. I also stated that his name will probably be mud. It will take some time to see what his future is in terms of AI research, but it certainly didn't talk long for him to be bounced from OpenAI.

The article goes on to cite WSJ and Bloomberg saying that "some new details emerged about the days leading up to Altman's firing. "In the weeks leading up to his shocking ouster from OpenAI, Sam Altman was actively working to raise billions from some of the world's largest investors for a new chip venture," Bloomberg reported. Altman reportedly was traveling in the Middle East to raise money for "an AI-focused chip company" that would compete against Nvidia."

To me, this seems like a good investment. I can't name anyone outside of Nvidia who makes chips for AI outfits, I'm sure there are a couple. Having another company designing and fabricating chips could be an excellent investment, it would break up the market which would stimulate competition (theoretically) and create better products. It would also increase availability - as I understand it, these chips are spoken for several years out. One way to get ahold of them is to wait for a current AI company go to bust and try to snap up their hardware at auction. You might be getting slightly older stuff, but at least you have something in-hand now.

Continuing, "As Bloomberg wrote, "The board and Altman had differences of opinion on AI safety, the speed of development of the technology and the commercialization of the company, according to a person familiar with the matter. Altman's ambitions and side ventures added complexity to an already strained relationship with the board."

OpenAI has an unusual structure, with a nonprofit that controls the for-profit subsidiary OpenAI Global. A Wall Street Journal behind-the-scenes report noted that the nonprofit board's mission is to "ensur[e] the company develops AI for humanity's benefit—even if that means wiping out its investors."

"According to people familiar with the board's thinking, members had grown so untrusting of Altman that they felt it necessary to double-check nearly everything he told them," the WSJ report said. The sources said it wasn't a single incident that led to the firing, "but a consistent, slow erosion of trust over time that made them increasingly uneasy," the WSJ article said. "Also complicating matters were Altman's mounting list of outside AI-related ventures, which raised questions for the board about how OpenAI's technology or intellectual property could be used."

Apparently the board was fine with OpenAI collapsing? Were that to happen, they would have very limited control on how their intellectual property got sold forward. Would they think that if Elon Musk were to sweep in and scoop it up and it would be a better situation?

At least the drama Altman and the Board is over. 75% of the board is gone, Altman isn't. More board members are going to be added, presumably to make the board less reactionary and more stable.


I haven't spoken a lot of my views of AI. They're not that complicated. AI is a disruptive technology. Yes, jobs are going to be lost, this is already proving true. How many more? Hard to say. I think it's too soon to say how much of a boon or threat this will be to people. The advent of the automobile was equally disruptive, it crushed the buggy and buggy whip makers. Same thing is going to happen here over time. Some jobs will be eliminated, other jobs will be created.

Will AI take over the world and quash humans entirely? There's an old saying, a computer's attention span is only as long as its power cord. AI requires huge computers/data centers to run. As long as we're not stupid enough to break the air gap in weapons systems and let AI completely run military systems, a series of bomber strikes will take out AI with little difficulty. The chips that run ChatGPT are leagues beyond what is in your PC or phone, it can't scurry off and hide, it needs huge data centers.

I'm not terribly worried about it. I've worked with ChatGPT 3.5, the free version, a fair amount. It's a useful assistant, but it won't replace people across the board.

https://arstechnica.com/tech-policy/2023/11/sam-altman-wins-power-struggle-returns-to-openai-with-new-board/

It was obvious the fallout from Friday's surprise sacking of Sam Altman wasn't remotely done, the big question was where he was going to land. I didn't think he was likely to begin a new startup because the length of time for one to become productive and competitive is too long, and the 'making a new AI chip' also has too long a payoff. Much more likely for him to land with another AI company.

And it turns out said AI company is Microsoft, a heavy investor in his former gig, OpenAI.

Altman will be heading up a "new advanced AI research team". And basically MS extended an open letter for anyone from OpenAI to join Sam in this new division - more on this. It seems to me that this will invite lawsuits from OpenAI about poaching talent, but they're kind doomed to fail. OpenAI has absolutely destroyed their market value with firing Altman, now might be a good time to short their stock as it's going to plumet.

https://techcrunch.com/2023/11/20/openai-co-founders-sam-altman-and-greg-brockman-to-join-microsoft/


Over 600 employees signed an open letter to the OpenAI board saying that if the ENTIRE board doesn't resign over the firing of Altman, THEY would resign. With Microsoft having picked up Altman, along with his co-founding buddy and perhaps everyone who resigned from OpenAI along with him, it looks like a pretty easy move for most people.

Over the weekend, "dozens" of OpenAI employees tweeted "OpenAI is nothing without its people". I look at it this way: OpenAI will be able to save a lot of money by right-sizing their office space needs and selling a lot of empty desks!

Now, here's the kicker:
“The process through which you terminated Sam Altman and removed Greg Brockman from the board has jeopardized all of this work and undermined our mission and company,” the letter reads. “Your conduct has made it clear you did not have the competence to oversee OpenAI.”

Remarkably, the letter’s signees include Ilya Sutskever, the company’s chief scientist and a member of its board, who has been blamed for coordinating the boardroom coup against Altman in the first place.

Shortly before the letter was released, Sutskever posted on X: “I deeply regret my participation in the board’s actions. I never intended to harm OpenAI. I love everything we’ve built together and I will do everything I can to reunite the company.” (emphasis mine)

It would appear that Sutskever seems to be having some second thoughts at seeing the potential looming collapse of OpenAI due to his and the board's actions.

Some more OpenAI follies: the board appointed Mira Murati, OpenAI’s chief technology officer ... as interim CEO. then after Altman et al was announced as joining Microsoft, "OpenAI’s board chose to remove Murati and appoint another interim CEO, Emmett Shear, the former CEO of Twitch, the video game streaming site."

Let's take a moment to process this. You fire the co-founder of the most recognizable AI company in the world. You replace him with your CTO for the interim CEO. Okay, I can dig that, at least Mira is probably a technologist, kinda goes with the title. And Mira probably hasn't had a chance to move into the new office and is yanked for THE FORMER CEO OF A VIDEO GAME STREAMING SITE?! As they say on Sesame Street (or used to, I haven't seen it in eons), 'One of these things is not like the other, one of these things just doesn't belong'.

I can just picture the board room discussion. 'Yeah, I know this dude who's looking for work right now, he'd be just perfect! He was the boss of this big rockin' company!'

Shear, according to his Wikipedia page, has zero background in data science or AI. He's a financial wonk from Yale.

Let's look at one last thing. Let's suppose those "over 600" OpenAI employees walk over to Microsoft.

OpenAI has 770 employees.

This is going to be a tremendous boon to Microsoft's AI development, they also have ties to chip makers, so Sam can pursue both sides.

https://www.wired.com/story/openai-staff-walk-protest-sam-altman/

You may remember a few months ago I posted a link to an article about a new AI supercomputer that consumed an insane amount of electricity, enough to power something on the order of 3,000 to 30,000 houses?

As you may suspect, consuming that amount of electricity requires A LOT of cooling. OpenAI has a datacenter that pulls from the watershed of the Raccoon and Des Moines rivers in central Iowa. Iowa. You know, cornfields. Breadbasket of America. I'd love to know how much water - precisely - they are pulling.

A researcher at UC Riverside "...estimates ChatGPT gulps up 500 milliliters of water (close to what's in a 16-ounce water bottle) every time you ask it a series of between 5 to 50 prompts or questions...

Google reported a 20% growth in water use in the same period, which Ren also largely attributes to its AI work.

OpenAI and Microsoft both said they were working on improving "efficiencies" of their AI model-training."

https://apnews.com/article/chatgpt-gpt4-iowa-ai-water-consumption-microsoft-f551fde98083d17a7e8d904f8be822c4

https://news.slashdot.org/story/23/09/10/2033253/to-build-their-ai-tech-microsoft-and-google-are-using-a-lot-of-water