It’s simple: you kill your old AI and train a new one. (Pray your AI doesn’t fight back.)
https://www.bespacific.com/they-need-to-be-entirely-rebuilt-every-time-theyre-updated/
They Need to Be Entirely Rebuilt Every Time They’re Updated
The Byte – “A new study highlights a glaring hole in AI models’ ability to learn new information: turns out, they can’t! According to the study, conducted by a team of scientists at Canada’s University of Alberta and published this week in the journal Nature, AI algorithms trained via deep learning — in short, AI models like large language models built by finding patterns in heaps of data — fail to work in “continual learning settings,” or when new concepts are introduced to a model’s existing training. In other words, if you want to teach an existing deep learning model something new, you’ll likely have to retrain it from the ground up — otherwise, according to the research, the artificial neurons in their proverbial minds will sink to a value of zero. This results in a loss of “plasticity,” or their ability to learn at all. “If you think of it like your brain, then it’ll be like 90 percent of the neurons are dead,” University of Alberta computer scientist and lead study author Shibhansh Dohare told New Scientist. “There’s just not enough left for you to learn.” And training advanced AI models, as the researchers point out, is a cumbersome and wildly expensive process — making this a major financial obstacle for AI companies, which burn through a ton of cash as it is…”
Perhaps we should use AI to read the reports, looking for errors?
Police officers are starting to use AI chatbots to write crime reports. Will they hold up in court?
… Oklahoma City’s police department is one of a handful to experiment with AI chatbots to produce the first drafts of incident reports. Police officers who’ve tried it are enthused about the time-saving technology, while some prosecutors, police watchdogs and legal scholars have concerns about how it could alter a fundamental document in the criminal justice system that plays a role in who gets prosecuted or imprisoned.
Built with the same technology as ChatGPT and sold by Axon, best known for developing the Taser and as the dominant U.S. supplier of body cameras, it could become what Gilbert describes as another “game changer” for police work.
… in Fort Collins, Colorado, where police Sgt. Robert Younger said officers are free to use it on any type of report, though they discovered it doesn’t work well on patrols of the city’s downtown bar district because of an “overwhelming amount of noise.”
(Related) Are prompts included as evidence?
https://www.makeuseof.com/ai-prompting-formula-guaranteed-results/
Try This AI Prompting Formula and I Guarantee You'll Love the Results
… Getting the result you want in one prompt often hinges on how you phrase your prompts. It’s like a conversation—you need to be clear, direct, and intentional about what you want. Here’s a simple formula that changed everything for me.
The Four-Part Prompting Formula
I wish I could take full credit for “inventing” this formula. However, I didn’t. This formula is a result of personal experience and my takeaways from Google’s 45-page prompting guide for Gemini.
Here’s the formula: Persona + Task + Context + Output Format. Let's break these down and see how they work together.
Important and likely still relevant.
NSA releases copy of internal lecture delivered by computing giant Rear Adm. Grace Hopper
FORT MEADE, Md. — “In one of the more unique public proactive transparency record releases for the National Security Agency (NSA) to date, NSA has released a digital copy of a lecture that then-Capt. Grace Hopper gave agency employees on August 19, 1982. The lecture, “Future Possibilities: Data, Hardware, Software, and People,” features Capt. Hopper discussing some of the potential future challenges of protecting information. She also provided valuable insight on leadership and her experiences breaking barriers in the fields of computer science and mathematics. Rear Adm. Hopper was an American computer scientist, mathematician, and United States Navy rear admiral. One of the first programmers of the Harvard Mark I computer, she was a pioneer of computer programming. Hopper was the first to devise the theory of machine-independent programming languages, and the FLOW-MATIC programming language she created using this theory was later extended to create COBOL, an early high-level programming language still in use today. In 2016, President Obama posthumously awarded Rear Adm. Hopper the Presidential Medal of Freedom — the Nation’s highest civilian honor, awarded to individuals who have made especially meritorious contributions to the security or national interest of the U.S. — for her remarkable influence on the field of computer science. While NSA did not possess the equipment required to access the footage from the media format in which it was preserved, NSA deemed the footage to be of significant public interest and requested assistance from the National Archives and Records Administration (NARA) to retrieve the footage. NARA’s Special Media Department was able to retrieve the footage contained on two 1’ APEX tapes and transferred the footage to NSA to be reviewed for public release. NSA recognizes Rear Adm. Hopper’s significant contributions as a trailblazing computer scientist and mathematician, but also as a leader.
“The most important thing I’ve accomplished, other than building the compiler, is training young people,” Rear Adm. Hopper once said. “They come to me, you know, and say, ‘Do you think we can do this?’ I say, ‘Try it.’ And I back ’em up. They need that. I keep track of them as they get older and I stir ’em up at intervals so they don’t forget to take chances.” [h/t Mike Ravnitzky]
Rear Adm. Hopper’s lecture and other public interest declassifications and transparency releases can be found under Historical Releases on NSA’s Declassification & Transparency Initiatives page.
No comments:
Post a Comment