The internet favorite ChatGPT has a new competitor, and his name is… Claude? OpenAI rival Anthropic has a chatbot to compete with heavyweights like ChatGPT and Google Bard. Claude has just got a significant update to give it a leg up on the competition: a massive memory increase.
Memory is one of the most significant limits of most chatbots, and it has received little attention throughout the current wave of AI popularity. It determines an AI’s ‘context window,’ or the amount of information a chatbot can process in a single encounter before it begins to ‘forget’ knowledge from prior interactions.
Although this data is not technically measured in words, ChatGPT can manage around 3,000 words of text in a single chat with a user before losing sight of things. If you’re a paying ChatGPT Plus subscriber, you may have had access to the limited-edition complete model of GPT-4, which can monitor three to four times that much.
Claude – I adore that name, by the way – can now comprehend about 75,000 words in a single encounter, thanks to a recent memory upgrade. Anthropic claims to be able to analyze a full book in less than 22 seconds; the business had Claude read The Great Gatsby again, with a single phrase altered – and the chatbot picked up on the difference in only 22 seconds.
A larger memory may result in fewer AI “hallucinations.”
This is a significant advancement in AI research since it would take the typical human five hours or more to read a work like The Great Gatsby. If you ask ChatGPT to explain a book, it will compile its answer from human-authored summaries across the internet; Claude might even create one himself. Regarding AI-assisted writing, it’s feasible that Claude will outperform ChatGPT.
Furthermore, memory is an unacknowledged issue for many AI chatbots right now. It’s why some chatbots, such as Microsoft’s Bing AI, have answer limitations to prevent the AI from acting weirdly when it ‘forgets’ prior points in a protracted discussion.
These conversational flaws are categorized as ‘AI hallucinations,’ which may result in chatbots creating incorrect replies to requests and acting strangely overall. Claude’s significantly expanded memory capacity will let it handle bigger chunks of knowledge and help it remain on topic throughout extended exchanges.
Anthropic improvеd AI isn’t yеt availablе to thе gеnеral public – you havе to bе a businеss partner with accеss to Claudе API, and thеrе’s no word on pricing – but it’s a stеp in thе right dirеction for improving chatbot capabilitiеs, еspеcially givеn how quickly thеy’rе bеing intеgratеd into еvеry aspеct of our digital livеs.
Of course, this does not address the many other issues with AI. It’s encouraging that Anthropic prioritizes safe and ethical AI research, declaring on its website that ‘we do not know how to teach computers to robustly act nicely’ – since no one does right now. But maybe we’ll get there, and Claude will be the AI that guides us there.