The open-source AI race has heated up stateside, just days after China’s DeepSeek stunned Silicon Valley with its shoestring-budget models.

The Allen Institute for AI (AI2) recently released a new model it claims can beat or hold its own with DeepSeek V3 and OpenAI’s GPT-4o across several benchmarks. The release is a larger version of the model that the nonprofit lab unveiled last November with nearly six times the number of parameters (405 billion vs. 70 billion).

At a moment when DeepSeek has thrust open-source AI into the spotlight, AI2 hopes that the model will demonstrate that US-based open-source companies are also chipping away at the performance gap between open and closed models.

“We only have a really, really limited number of open-source US-based models, like our model, and also [Meta’s] Llama and maybe a couple more,” Hannaneh Hajishirzi, senior director of NLP research at AI2, told Tech Brew. “So we showed that applying [our training regimen] on a US open language model, you can actually close the gap, or even outperform DeepSeek V3.”

Keep reading here.—PK

The City of Love was not all bonhomie last week as world leaders descended on Paris for a sometimes-contentious summit on the future of international AI development.

Vice President JD Vance set the tone for the Trump administration’s new approach to AI safety with a speech that laid out a vision of US dominance of the tech and criticized European digital regulations. The US and the UK also declined to sign a non-binding pledge calling for more “inclusive and sustainable” AI development backed by more than 60 countries.

The summit also comes as seemingly ultra-efficient generative AI models from DeepSeek and other Chinese companies have intensified a global arms race with the US around the technology.

Safety off: Vance purposefully sought to mark a contrast from the first of these annual summits in England in 2023, when 28 governments—including the US and the UK—signed onto an agreement that noted risks of “serious, even catastrophic, harm” from AI models.

“I’m not here this morning to talk about AI safety, which was the title of the conference a couple of years ago. I’m here to talk about AI opportunity,” Vance said. “The AI future is not going to be won by hand-wringing about safety. It will be won by building.”

Keep reading here.—PK

Clean up your mess.

That’s the message from Lead the Charge, a coalition of advocacy organizations pushing for “an equitable, sustainable, and fossil-free auto supply chain,” to automakers around the world.

On Feb. 11, the group released its third annual Auto Supply Chain Leaderboard, which ranks 18 global automakers based on their efforts to rid their supply chains of “emissions, environmental harms, and human rights violations” amid the industry’s transition to electric vehicles. While EVs don’t produce tailpipe emissions, their supply chains aren’t always clean.

This year’s report concludes that “despite notable achievements in some areas by automakers, the industry overall is making slow and inadequate progress on cleaning up its supply chains: For the third year running, no automaker achieved a total score of over 50% and the total average score across all automakers is just 22%.”

Keep reading here.—JG