Apr 3, 03:19 PM

OpenAI's Sam Altman Announces Open-Weight Model Amid $40B Funding, Rivalry with DeepSeek

OpenAI has announced plans to release its first open-weight language model since GPT-2, responding to competitive pressures from rivals such as DeepSeek and Meta's Llama 3. This model will allow developers to access and modify trained parameters without revealing the full source code or training data, offering more flexibility in AI development. CEO Sam Altman has acknowledged the need for a shift towards more open models, citing the success of DeepSeek's R1 model as a catalyst for this strategic change. The move comes as OpenAI navigates the challenges of an AI industry facing infrastructure limitations, with the company recently securing a $40 billion funding round to reach a $330 billion valuation. In the broader AI landscape, startups like Mercor are achieving significant revenue milestones, with Mercor reaching $100 million in annualized revenue run rate (ARR) in just 11 months. However, the sustainability of these revenue figures is under scrutiny, as 60-70% of Mercor's revenue is paid out to contractors, and the company was last valued at $2 billion. Meanwhile, Sentient, backed by Peter Thiel's Founders Fund, has launched Open Deep Search (ODS), an open-source AI search system designed to outperform competitors like Perplexity and ChatGPT. Sentient Chat, its AI assistant, has already attracted over 1.75 million users to its waitlist.