I call this “real data.” Data that has not been contaminated by AI generated nonsense. (It has been contaminated by disinformation and other random noise.) Does this suggest that what exists today is a good as something like ChatGPT will ever get?
https://www.businessinsider.com/ai-could-run-out-text-train-chatbots-chatgpt-llm-2023-7
Generative AI tools are quickly 'running out of text' to train themselves on, UC Berkeley professor warns
ChatGPT and other AI-powered bots may soon be "running out of text in the universe" that trains them to know what to say, an artificial intelligence expert and professor at the University of California, Berkeley says.
Stuart Russell said that the technology that hoovers up mountains of text to train artificial intelligence bots like ChatGPT is "starting to hit a brick wall." In other words, there's only so much digital text for these bots to ingest, he told an interviewer last week from the International Telecommunication Union, a UN communications agency.
… A study conducted last November by Epoch, a group of AI researchers, estimated that machine learning datasets will likely deplete all "high-quality language data" before 2026. Language data in "high-quality" sets comes from sources such as "books, news articles, scientific papers, Wikipedia, and filtered web content," according to the study.
(Related)
https://www.bespacific.com/a-categorical-archive-of-chatgpt-failures/
A Categorical Archive of ChatGPT Failures
A Categorical Archive of ChatGPT Failures Ali Borji. Quintic AI. April 5, 2023 “Large language models have been demonstrated to be valuable in different fields. ChatGPT, developed by OpenAI, has been trained using massive amounts of data and simulates human conversation by comprehending context and generating appropriate responses. It has garnered significant attention due to its ability to effectively answer a broad range of human inquiries, with fluent and comprehensive answers surpassing prior public chatbots in both security and usefulness. However, a comprehensive analysis of ChatGPT’s failures is lacking, which is the focus of this study. Eleven categories of failures, including reasoning, factual errors, math, coding, and bias, are presented and discussed. The risks, limitations, and societal implications of ChatGPT are also highlighted. The goal of this study is to assist researchers and developers in enhancing future language models and chatbots. Please refer to here for the list of questions.”
I wonder if this is the only app the government deploys?
Watch: Government Spyware on Your Phone? Unfortunately, There’s an App for That
Washington, DC (July 9, 2023) – The New Civil Liberties Alliance is challenging the Massachusetts Department of Public Health (DPH) in federal court for coordinating with Google to automatically install spyware on the smartphones of more than one million Commonwealth residents, without their knowledge or consent, in a misguided effort to combat Covid-19. A newly-released video details how DPH’s actions have violated fundamental constitutional rights.
Thousands of people do not know DPH’s Covid-19 tracking app is on their phone, as it does not appear on their home screens like other apps. NCLA client Robert Wright, who commutes to Massachusetts for work, was appalled to learn that the government put an app on his phone without his knowledge, especially one that could constantly track his movements. NCLA’s lawsuit argues the DPH app’s automatic installation infringes on the Fourth Amendment right to privacy because it interferes with phone owners’ private property and collects information about them. By taking up storage space on phones against their owners’ will, such unwanted installations also constitute uncompensated taking of property in violation of the Fifth Amendment.
Humorous or truly scary?
Budding politician ‘has no policies,’ will use AI to legislate
… “If elected, I will vote in Parliamant [sic] in accordance with the consensus. Simple,” Gray wrote on LinkedIn.
The tool he’s using, Polis, collects and analyses public opinion in real time.
I’m surprised it took so long…
https://cointelegraph.com/news/google-hit-with-lawsuit-over-ai-privacy-policy
Google hit with lawsuit over new AI data scraping privacy policy
A week after Google updated its privacy policy to allow data scraping for AI training purposes, the company faces a class-action lawsuit.
Tools & Techniques.
https://www.makeuseof.com/build-custom-chatgpt-with-your-own-data/
How to Build a Custom ChatGPT With Your Own Data
Looking to provide ChatGPT with your custom data? Here's a step-by-step on how to do just that!
Tools & Techniques.
Anthropic — the $4.1 billion OpenAI rival — debuts new A.I. chatbot and opens it to public
There’s a new entrant in the budding AI arms race.
As Microsoft -backed OpenAI and Google race to develop the most advanced chatbots, powered by generative artificial intelligence, Anthropic is investing heavily to keep up. Just a few months after raising $750 million over two financing rounds, the startup is debuting a new AI chatbot: Claude 2.
Founded in 2021 by former OpenAI research executives and funded by companies including Google, Salesforce and Zoom, Anthropic is opening up its chatbot technology to consumers for the first time with Claude 2. For the past two months, the company’s AI models have been tested by businesses such as Slack, Notion and Quora, and Anthropic has accumulated a waitlist of more than 350,000 people requesting access to Claude’s application programming interface and its consumer offering.
No comments:
Post a Comment