—Michael Lyles, B1Daily

The 2021 San Francisco tech startup, Anthropic, released its patented flagship artificial intelligence system, nicknamed Claude. 

Anthropic is headed up by founders Dario and Daniela Amodei, who are CEO and President. Both have extensive ties in Silicon Valley. The paid rubbed competitors the wrong way when GPT-3 engineer Tom Brown and fellow Co-founder Sam McCandlish left their company to join Anthropic. They also boast over 20 Open AI engineers who switched to their company. 

In 2023, Anthropic partnered with Amazon’s WorksSpace (AWS) utilizing their high powered African-mined Trainium and Inferentia chips to build, train, and deploy its future foundation models at a speed only second to DeepSeek AI.

Amazon donated over $4 billion in revenue and since the company has been on a spending spree, developing 3 different updates for the Claude artificial intelligence agent.

Claude expanded its “context window” from 9,000 tokens to 100,000 tokens in May 2023 shortly before releasing Claude 2. This means it can handle inputs of around 75,000 words, so businesses can submit huge reams of documentation for Claude to analyze. But Claude 3’s family has an even bigger context window, and shown very respectable performance and has a longer (200k) context windows across the entire family (Opus + Sonnet + Haiku).

Anthropic said its latest artificial intelligence model resorted to blackmail when told it would be taken offline.

In a safety test, the AI company asked Claude Opus 4 to act as an assistant to a fictional company, but then gave it access to (also fictitious) emails saying that it would be replaced, and also that the engineer behind the decision was cheating on his wife. 

Anthropic said the model “threatened to reveal the affair” if the replacement went ahead.

In conclusion, Antrhopic’s AI agent is proving to be far better at prolonged ponderance than we could have ever hoped for. 

—Michael Lyles, B1Daily