Anthropic's chatbot Claude AI arrives in Italy
Claude's arrival in Europe comes as a result of additional work to comply with the stringent regulations of the AI Act, effective next year.
3' min read
3' min read
Italian-American siblings Dario and Daniela Amodei are finally bringing their 'artificial intelligence with humans at the centre' to Europe. Unlike their competitors, ChatGPT - which launched the free GPT-4o model a few days ago - and Google's Ai Gemini, Claude's founders claim a different approach to development: Helpfull, Honest, Harmless. A business model, which, according to the founders, wants to make ethics and competitiveness work together, and which has convinced important financiers, including Google and Amazon.
Three years after the start of the adventure and two months after its launch, Anthropic has announced in an official post the landing of the Claude platform (Claude.ai) for European users, mainly companies, but with the intention of also trying to enter people's daily habits, thanks to the iPhone app already available and the Android app (coming soon). According to one's needs, one can choose between the three available models: Claude 3 Haiku, Claude 3 Sonnet and Claude 3 Opus, the most powerful of the lot.
What Claude 3 promises to do
Compared to previous generations, the improvements in the third version of the chatbot are significant. In its presentation, Anthropic emphasised that it had achieved greater accuracy, with less frequent errors and a better ability to understand and answer complex and factual questions, achieving a negligible rejection and rejection rate of requests.
The pro version, according to the developers, demonstrates 'near-human levels of understanding and fluency in complex tasks' and promises to raise the bar in a wide range of cognitive tasks, outperforming its competitors in most of the evaluation benchmarks used to assess artificial intelligence systems, including expert knowledge at undergraduate level (Mmlu), expert reasoning at graduate level (Gpqa) and basic mathematics (Gsm8k). Opus, in particular, the best performing model of the trio, is presented as an 'intelligence superior to any other model now available' and can be used to plan and execute complex actions on programming interfaces (APIs), interactive coding and databases. At the strategic level, it can also perform research reviews, brainstorming and generate hypotheses. It will also delight those who work with large numbers, thanks to advanced analysis of financial data, including charts and graphs, generating market trends and forecasting future scenarios. On top of all this, drug discovery is also possible.
