DeepSeek gives devs 'high-performance AI' on a budget, but it comes at a cost

Lorikeet News Desk

published

February 10, 2025

Credit: Outlever

Key Points

DeepSeek's launch prompts US AI leaders to release more affordable models.
Experts highlight the potential shift in global AI dominance from the US to China, causing the US to move towards faster model releases.
Calls for policy action emphasize infrastructure investment, privacy frameworks, and incentives for AI competitiveness.

Rather than focusing solely on restricting access to advanced hardware, US policy should incentivize local development of more efficient architectures and secure deployment models to ensure the US remains globally competitive.

Jonathan Baker

Principal Cloud Engineer | BlackLine

Markets tumbled and American AI leaders scrambled to release cheaper models following DeepSeek's launch. Tension grips the US tech sector as it faces the possibility of losing its position as the world's AI leader to China.

We spoke with two AI experts—Jonathan Baker, Principal Cloud Engineer at Nasdaq-listed finance automation platform BlackLine, and Danny Hammo, Founder and CEO of ExecAI—about DeepSeek's launch's domino effect and potential US responses.

Global power shift: "US players were caught off guard," says Hammo. "They were the dominant players in general AI and advanced model capabilities—they were the 'future makers.' However, DeepSeek changed the dominance equation. This new 'threat' will likely pressure US AI companies to accelerate their release cadence and perhaps launch a new generation of models earlier than planned."

Democratizing AI development: Looking ahead to 2025, Hammo predicts a surge of VC-backed startups, inspired by DeepSeek's demonstration that high-performance AI doesn't require astronomical budgets. "DeepSeek has shown that the barrier to entry is lower than imagined," he notes, suggesting the dawn of a new era in accessible, competitive AI development.

US players were caught off guard. They were the dominant players in general AI and advanced model capabilities—they were the 'future makers.' However, DeepSeek changed the dominance equation.

Danny Hammo

Founder and CEO | ExecAI

Policy Imperatives: Baker calls for a proactive US policy response to maintain AI competitiveness: "The US needs specific and actionable policies to maintain competitiveness while protecting national interests. A comprehensive approach should start with infrastructure investment, creating tax incentives for companies investing in AI infrastructure, coupled with funding for research into efficient AI architectures and support for domestic chip manufacturing focused on AI-specific hardware."

Privacy Framework: "Clear data protection frameworks must establish requirements for AI companies to disclose training and inference locations, data collection practices, and storage locations," Baker explains. "A certification framework, similar to SOC 2 compliance, would help consumers evaluate AI models' privacy practices. The use of AI in critical applications such as defense and healthcare should be regulated and tested under supervision."

Baker concludes: "Rather than focusing solely on restricting access to advanced hardware, US policy should incentivize local development of more efficient architectures and secure deployment models to ensure the US remains globally competitive."

One more thing