|
|
From Opalesque Geneva: After comparing the AI frenzy to the dot-com bubble in a paper last year, portfolio manager Lukasz Tomicki is changing his mind.
He has recently been impressed by the latest models from OpenAI and DeepSeek. "These models have moved beyond mere regurgitation of training data," he says in his latest monthly report seen by Opalesque. "They're showing real depth in reasoning and problem-solving."
Tomicki, who manages a systematic long/short hedge fund, has been using DeepSeek-V3 for months and found this tool to be very useful for software development and language translation, especially with Python and C++. "That said, I don't believe DeepSeek (or any LLM) replaces a programmer. The code it generates is a solid starting point, but it still needs debugging and integration. Where it truly shines, though, is in helping me understand unfamiliar code."
The introduction of Chain of Thought (CoT) reasoning, sometimes called "deep thinking," was a breakthrough that started with OpenAI last year and continued with DeepSeekR1 in January. The latter is fully open-source, which "means anyone can build on its foundation, accelerating AI development in ways that proprietary models simply can't match."
DeepSeek-R1's API (application programming interface) costs about 1/200th of what OpenAI charges. Furthermore, DeepSeek uses a mixture-of-experts (MoE) architecture, meaning only a small portion of the model is activated for any give...................... To view our full article Click here
|
|
RSS
