LyrArc Article Gist
Doing a 2 hour walk in Sydney, Australia, using ChatGPT turns out to be not so much fun with inaccurate replies casually given, and distances being off, restaurants closed and train stations nonexistent, says travel expert Kate Hennessy in The Guardian.

LyrArc Article Gist
This report in The Guardian shows that ChatGPT is nothing new. The first version of this kind of generative AI was developed in 1966 at MIT by a computer scientist Weizenbaum, who called it Eliza. The buzz around it like that around ChatGPT was that it was thinking and acting on its own, the way humans like to think it did, but in fact Weizenbaum showed that it was simply code written to take what was given to the computer as input and spitting it out in a different way that made it look that it was acting on its own, when it clearly was nothing but parroting it out like a parrot. The issue of turning our world over to robots based on AI is controversial and even dangerous. A Japanese futuristic movie shows how the man who has written the code for the master computer that runs everything in Japan is disillusioned about it and finds himself in a nightmare world where the machine tries to isolate and eliminate the man who created it. Machines cannot think or have emotions like humans do and it is these emotions, rethinking, that the world depends on for its survival. Can anyone say that a machine would have made the decision that Chinese president Jinping just made in January of making a complete u turn and moving away completely from lockdowns into a complete opening with a plan that appears to have worked and is reviving China's economy following the street protests by informal groups including young women? ...

LyrArc Article Gist
What Joanna Stern doesn't say in the title when she says she passed is that she got C grades and could not contribute much to the AP English class she joined at a Secaucus, N.J. high school class. Grammarly to check grammatical and spelling errors is one thing and English essay writing using AI tools is another. Most of all it is about not learning even when it can be difficult. And some literary works such as Kafka's are difficult for most of us.

LyrArc Article Gist
This report in WSJ says fears are growing about the potential misuse of artificial intelligence that includes committing crimes and spreading false information, or information that is likely to be damaging to people. The Biden administration is looking into the possible misuse and inviting comment. The National Telecommunication and Information Administration is as a first step inviting comments. Alan Davidson who heads the agency says we know we have to put some guardrails in place. Asked about AI and new chatbots and risks they pose president Biden says, "it remains to be seen," it could be.

LyrArc Article Gist
How ChatGPT and Microsoft are seen as upending the search business at Google. ChatGPT's ability to explain complex concepts and come up with new thoughts are seen as creating anew way of doing Search. Microsoft is investing billions of dollars in ChatGPT. Much of Google's efforts in the last five years have gone into preserving its monopoly of the search business without adding much through innovation and new ideas. The monopolistic behaviours have attracted the attention of regulators. Microsoft once a monopolistic behaviour company is now playing the role of a challenger with new ideas and concepts.

LyrArc Article Gist
Chinese bio tech hubs around Shanghai started by biotech engineers from China trained in the US returning to China are giving the US competition. These biotech companies in China are coming up with drugs for lung cancer and other fields at a fraction of the cost. This happened in AI with DeepSeek doing what ChatGPT does at 10% of the cost.

LyrArc Article Gist
Matt Garman, head of AI and Cloud Services at Amazon talks to Emma Tucker of the WSJ. Garman says Amazon was not slow, just being deliberate and thinking it through rather than coming up with something quickly as Microsoft had done with ChatGPT. He sees the need for regulation, only that it not create a situation where others including China go ahead and the US falls behind. This means that the US needs to coordinate AI rules with other countries including China, India, Russia, for comprehensive solutions on how AI is to be managed to work towards good.

On a five day week vs remote work Garman says Amazon takes the view that the creative work can best be done with humans interacting at the office. It sees this as essential for good work. On the 3 day week with 2 for remote work, the only problem he says is that everyone picks different 3 days and this leads to loss of human interaction at work.

LyrArc Article Gist
For two decades young computer specialists in their 30's have tried to say the right things to the American people to gain approval- with AI this is where it all ends. Watching Murthi and Altman on Stern's interview in WSJ one senses a lack of awareness that this is too big a technology to be handled by a few computerized work  specialists. Murthi says the right things with a faltering conviction, Altman lacks conviction, yet both cannot take on the responsibilities for AI on their shoulders. The chief technology officer of OpenAI takes over role of Interim CEO with the departure of Sam Altman. She graduated from the Thayer School of Engineering at Dartmouth. She worked at Tesla before joining OpenAI. Murthi was in charge of Operations, managing the teams that delivered ChatGPT product in 2022, and handling the relationship with Microsoft which invested $13 billion for a 49% stake in OpenAI.

LyrArc Article Gist
Anthropic's settlement for $1.5 billion at $3000 a book, and it's efforts in Education that conflict with the Nation's need to get 4th graders to Read and Learn. Anthropic's website shows it trying to get into Education and to measure the Economic Index from effects of AI. Yet the pretensions to goodwill for the public cause is not supported by facts, facts that the AI companies have nothing to show for the dismal situation for Global Literacy that is the case today. Literacy in the US that is dismal with about two thirds of 4th graders not able to read and comprehend the English language at a level of proficiency in American schools. These AI purveyors care only for the money they can make using vast amounts of electricity for these servers, and pretensions for public purpose are intended to smooth their access to public resources not some genuine interest in whether kids can read, which requires the hard work of the teachers in the public and private schools of this Nation and others in Europe, Asia, Latin America and Africa. The Movement for Global Literacy is Lyrarc's effort to support reading and learning and Lyrarc serves this purpose without such massive funding and without charging for the public service to the Nation and to other Nations in the world community. Anthropic settlement of $1.5 billion at $3000 a book for its AI bots use of copyrighted books, can lead to future litigation for OpenAI model that consumes vast amounts of data. Anthropic was founded by siblings Daniela and Dario Amodei after leaving OpenAI in 2021 in San Francisco. It hired Google Books Turvey to scan books for its large language models on a massive scale to train Claude its version of OpenAI's ChatGPT.  An investment of $4 billion by Amazon and additional $2 billion by Google provided funding. In this way it is a competitor to Microsoft funded OpenAI which made early advances in AI.  This article in WSJ says by making the settlement for $1.5 billion Anthropic is trying to make it harder for Open AI to scan material easily without paying for the access and thus blocking it's rival.   ...

LyrArc Article Gist
Surges in capital value can be wildly misleading. Nvidia a rapid computing company propelled in stock value. From the growth of crypto currency that led to losses and was perceived as a danger to the financial system by central banks and governments. This is happening when capital investment is a dire need in education and schools, good teachers and good classrooms, when only a third of American students pass NAEP tests on reading comprehension. Today's capital allocation system was never designed to accomplish this even as it sends hundreds of billions of dollars in one single day to a single company. Nvidia is now seeing a surge from chatbots computing coming out of ChatGPT,  leading to $184 billion change in its market value on May 25, 2023.  Nvidia was mostly a graphics processing company setup to make graphics on PC's look better. In 2006 Jensen Huang made the decision to open it up to developers to tinker with it and develop more computing capabilities. This has led to Nvidia designing much more powerful computing chips that perform thousands of calculations at the same time.   Nvidia designs the chips and sends production out to Taiwan Semiconductor Manufacturing. Suddenly Nvidia sees its share price surge and it joins companies such as Apple, Amazon, Microsoft, and Tesla that have seen one day surge in the value of the companies by over $100 billion shown in this WSJ graph by date. Huang says he thinks that this is the beginning of a ten year period in which companies will redo their data centers to build them up with AI computing capabilities. WSJ also says China's top nuclear weapons research institute has bought these advanced chips even though it is on a US export blacklist since 1997. In 2022 the Biden administration imposed new licensing requirements on export of the most advanced chips. Since then Nvidia is following specifications for chips that allow it to export to China, says the WSJ.     ...