A recap of our CES 2025 coverage; Anthropic reportedly looking to raise funds; data centers' impact on the elecric grid; Mistral AI updated Codestral; xAI's Aurora unsurprisingly has no guardrails; Raspberry AI's Series A; ChatGPT's new beta feature; LetsData's fundraise.
LetsData, a cybersecurity startup born during Russia's invasion of Ukraine, has raised $1.6M to expand its AI-powered information operations detection technology that analyzes content across multiple platforms and languages.
OpenAI has introduced a new "tasks" feature in beta for ChatGPT's paid users that enables scheduling reminders and recurring requests, marking the company's first step into AI agent capabilities.
Fashion design platform Raspberry AI has secured $24 million in Series A funding led by Andreessen Horowitz to expand its AI technology that converts fashion sketches into instant photorealistic renderings, already used by major brands like Under Armour and MCM Worldwide.
X users continue to take advantage of Grok's lack of guardrails to generate and flood the platform with harmful content. Reports point to a spike in the prevalence of abusive content after Grok's image generation capabilities were enhanced with Aurora, xAI's proprietary model.
Mistral recently announced an update to its code generation model, equipping it with a new architecture and improved tokenizer. Codestral 25.01 supports a leading 256K context length, performs strongly on popular coding benchmarks, and outperforms its rivals on fill-in-the-middle (FIM) tasks.
A Bloomberg analysis reveals that homes within 20 miles of AI data centers are experiencing dangerous levels of power grid distortions (bad harmonics), particularly in Virginia's "data center alley". Meanwhile, tech companies plan massive expansions of their data center infrastructure across the US.
After securing a $4B investment from Amazon in November 2024, Anthropic is reportedly close to raising an additional $2B led by Lightspeed Venture Partners, which would value the AI company at $60B.
At CES 2025, Nvidia unveiled its GeForce RTX 50 Series graphics cards, headlined by the $2,000 RTX 5090 with 21,760 CUDA cores and 32GB GDDR7 memory, promising twice the performance of previous generations in both gaming and AI applications.
AMD announced its new Ryzen AI Max processor series at CES 2025, featuring up to 16 Zen 5 CPU cores, RDNA 3.5 graphics, and enhanced AI processing capabilities. The company also added more models to its Ryzen AI 300/300 PRO Series and introduced the more affordable 200/200 PRO Series.
DeepSeek, a Chinese AI firm backed by quantitative hedge fund High-Flyer, recently released DeepSeek V3, a model available for commercial and non-commercial applications. DeepSeek V3 displays competitive performance that approaches recent models and is particularly proficient in coding.
Google contractors have been reportedly asked to compare and score Gemini's responses against Claude's using criteria including truthfulness and verbosity. A spokesperson for Google DeepMind confirmed the company was only using Claude for model output comparison, a standard industry practice.