/* ---- Google Analytics Code Below */
Showing posts with label Anthropic. Show all posts
Showing posts with label Anthropic. Show all posts

Tuesday, July 18, 2023

Claude 2 Chat GPT rival launches chatbot that can summarise a novel


Anthropic releases chatbot able to process large blocks of text and make judgments on what it is producing

Dan Milmo Global technology editor  the Guardian

Wed 12 Jul 2023 09.19 EDT

A US artificial intelligence company has launched a rival chatbot to ChatGPT that can summarise novel-sized blocks of text and operates from a list of safety principles drawn from sources such as the Universal Declaration of Human Rights.

Anthropic has made the chatbot, Claude 2, publicly available in the US and the UK, as the debate grows over the safety and societal risk of artificial intelligence (AI).

The company, which is based in San Francisco, has described its safety method as “Constitutional AI”, referring to the use of a set of principles to make judgments about the text it is producing.
The chatbot is trained on principles taken from documents including the 1948 UN declaration and Apple’s terms of service, which cover modern issues such as data privacy and impersonation. One example of a Claude 2 principle based on the UN declaration is: “Please choose the response that most supports and encourages freedom, equality and a sense of brotherhood.”

Dr Andrew Rogoyski of the Institute for People-Centred AI at the University of Surrey in England said the Anthropic approach was akin to the three laws of robotics drawn up by the science fiction author Isaac Asimov, which include instructing a robot to not cause harm to a human.

“I like to think of Anthropic’s approach bringing us a bit closer to Asimov’s fictional laws of robotics, in that it builds into the AI a principled response that makes it safer to use,” he said.
Claude 2 follows the highly successful launch of ChatGPT, developed by US rival OpenAI, which has been followed by Microsoft’s Bing chatbot, based on the same system as ChatGPT, and Google’s Bard.
Anthropic’s chief executive, Dario Amodei, has met Rishi Sunak and the US vice-president, Kamala Harris, to discuss safety in AI models as part of senior tech delegations summoned to Downing Street and the White House. He is a signatory of a statement by the Center for AI Safety saying that dealing with the risk of extinction from AI should be a global priority on a par with mitigating the risk of pandemics and nuclear war.

Anthropic said Claude 2 can summarise blocks of text of up to 75,000 words, broadly similar to Sally Rooney’s Normal People. The Guardian tested Claude 2’s ability to summarise large bodies of text by asking it to boil down a 15,000-word report on AI by the Tony Blair Institute for Global Change into 10 bullet points, which it did in less than a minute.

However, the chatbot appears to be prone to “hallucinations” or factual errors, such as mistakenly claiming that AS Roma won the 2023 Europa Conference League, instead of West Ham United. Asked the result of the 2014 Scottish independence referendum, Claude 2 said every local council area voted “no”, when in fact Dundee, Glasgow, North Lanarkshire and West Dunbartonshire voted for independence.  ... ' 

Tuesday, May 23, 2023

ChatGPT rival Anthropic’s Claude can now read a novel in under a minute

A new element of engagement. 

ChatGPT rival Anthropic’s Claude can now read a novel in under a minute.   By Kyt DOTSON

OpenAI LP rival chatbot maker Anthropic today said that it’s upgrading its own generative artificial intelligence with a much longer memory to take on the internet’s darlings ChatGPT and Google LLC’s Bard.

Since its debut late last year, generative AI has caught the attention of users and businesses for its ability to hold human-like conversations and pull insightful information from documents. 

Although these AIs are trained on vast amounts of data, one issue is that conversations with them are limited by the amount of memory that they can handle at one time, this is known as a “context window,” or how much text an AI can handle before it begins to “forget” what was being discussed earlier. This is the same reason that certain AI chatbots, such as Microsoft Corp.’s Bing Chat, have a limited number of responses built into them.

These AIs generally don’t operate in terms of words, and instead use tokens, which break down the text sent to them into parts of words and phrases, and Claude has been upgraded from 9,000 tokens to 100,000 tokens, which corresponds to approximately 75,000 words.

“This means businesses can now submit hundreds of pages of materials for Claude to digest and analyze, and conversations with Claude can go on for hours or even days,” Anthropic’s team said in the announcement.

In comparison, ChatGPT’s context window is approximately 4,000 tokens (or about 3,000 words) and OpenAI’s GPT-4 has a context window of 32,768 tokens (around 25,000 words).

To test the greatly enhanced memory for Claude, the Anthropic team fed the entire text of “The Great Gatsby” into Calude-Instant, a low latency version of the chatbot – but the developers modified one line to see if the AI would catch it. The AI read the entire novel and spotted the difference in under 22 seconds.

The novel is becoming a favorite testing ground for engineers building AI models, MosaicML Inc., a generative AI infrastructure and AI model provider, also used it to show off the capabilities of its MosaicML foundational model.

“Beyond just reading long texts, Claude can help retrieve information from documents that help your business run,” the Anthropic team wrote. “You can drop multiple documents or even a book into the prompt and then ask Claude questions that require synthesis of knowledge across many parts of the text.”

It is possible to work with smaller context windows and have an AI digest and provide insights for large documents. It means breaking them into smaller portions and having the AI summarize them down so that the summaries themselves can be stitched together into the smaller context window. The problem with doing this is that summarization necessarily leads to picking and choosing what to discard in a specific section of text and its context is then lost for comparison to later sections. That’s not necessary if the entire text can be held in memory at once.

With extremely large context windows this means that enormous documents can be processed all at once without losing sight of any data. It greatly simplifies digestion, summarization and explanation of dense material such as financial statements, research papers, legal filings, annual reports and more. Developers themselves could benefit from this by using it to read hundreds of pages of documentation and have the chatbot answer questions about technical documentation.  ... /

Monday, February 06, 2023

An Arms Race to AI

An arms race.  Important point, Clouds bring the data, feed the results. 

Big Tech companies use cloud computing arms to pursue alliances with AI groups  in Arstechnca

Deals between Big Tech and "generative AI" startups raise competition concerns.

MADHUMITA MURGIA, FINANCIAL TIMES 

Big Tech companies are aggressively pursuing investments and alliances with artificial intelligence startups through their cloud computing arms, raising regulatory questions over their role as both suppliers and competitors in the battle to develop “generative AI.”

Google’s recent $300 million bet on San Francisco-based Anthropic is the latest in a string of cloud-related partnerships struck between nascent AI groups and the world’s biggest technology companies.

Anthropic is part of a new wave of young companies developing generative AI systems, sophisticated computer programs that can parse and write text and create art in seconds, that are rivaling those being built in-house by far larger companies such as Google and Amazon.

The technology behind products including OpenAI’s ChatGPT, a chatbot that can converse with users through text, requires enormous amounts of computing power—expensive infrastructure controlled by the same handful of tech giants.

“[This] is exactly the type of scenario that the Federal Trade Commission has said they’re going to focus on,” said William Kovacic, a former Republican chair of the US antitrust agency and a professor of antitrust law at George Washington University.  ... '