They weren’t alone with announcements. DeepSeek has announced that its upgraded artificial intelligence model demonstrates improved capabilities in mathematics, programming, and logic while reducing instances of hallucination. This enhancement to the R1 model comes after it previously gained attention for competing with larger U.S. developers at a significantly lower cost. The upgrade reflects a deeper reasoning ability, which DeepSeek claims sets it apart from previous iterations. The announcement was made via a post on the AI model platform Hugging Face, emphasizing the advancements made despite the company’s limited resources compared to its competitors.
And notably, Reddit has filed a lawsuit against Anthropic accusing it of illegally using data from Reddit’s platform to train its AI systems. The lawsuit claims that Anthropic attempted to access Reddit data over 100,000 times, violating the site’s content policies and enriching itself without compensating Reddit or respecting user privacy. Reddit’s Chief Legal Officer, Ben Lee, stated that the company will not tolerate the commercial exploitation of its content without proper returns for its users. The lawsuit highlights the ongoing conflict between social media companies and AI firms over data usage, as companies increasingly restrict access to their data to protect user privacy. Reddit, which has over 100 million daily users, has recently begun to recognize the value of its user-generated content and is seeking licensing agreements with other companies in the industry.
ChatGPT continues to dominate the enterprise artificial intelligence landscape, according to New Relic’s first “AI Unwrapped: 2025 AI Impact Report,” which analyzes data from 85,000 businesses. The report reveals that over 86% of all large language model tokens processed by New Relic customers were associated with ChatGPT, highlighting its widespread adoption for general-purpose tasks. Chief Technical Strategist Nic Benders noted that while ChatGPT remains the leading model, developers are rapidly exploring new options as they are released. Notably, there has been a 92% increase in the number of unique models used in AI applications in early 2025, indicating a trend towards broader model experimentation. Python has reaffirmed its position as the favored programming language for AI, with a 45% increase in adoption since the last quarter, while Node.js and Java also saw significant growth. The report underscores the increasing need for robust AI monitoring solutions, with New Relic reporting a steady 30% growth in the use of its AI Monitoring service.
Why do we care?
ChatGPT isn’t going anywhere—but the dominance of one model is giving way to fit-for-purpose AI stacks, shaped by legal, financial, and performance constraints.
Reddit’s lawsuit against Anthropic underscores the increasing friction around data sourcing for AI. This has direct implications for IT providers offering AI-infused services, particularly those integrating user-generated content or third-party platforms:
- Legal exposure is real: Any AI solution that pulls data from the open web (or platforms like Reddit, StackOverflow, or Twitter) may be subject to retroactive licensing claims.
- Data provenance matters: IT providers must vet the datasets and model training pipelines of their vendors—especially in regulated or IP-sensitive sectors.
Smart providers will:
- Help clients navigate AI model selection like cloud choice—balancing price, capability, and compliance.
- Build practices around AI observability, monitoring, and governance, following the signal from New Relic’s growth in monitoring demand.
- Invest in developer talent and AI orchestration frameworks to deliver modular, adaptable solutions—not locked-in workflows.
In short: AI is no longer about picking the best model—it’s about delivering measurable value using the right model for the right job. That’s a core services opportunity.