Large Language Models

Today's Stories

1. 1

   Interactive Brokers has integrated ChatGPT and Grok into its trading platform, giving customers access to AI assistants for investment decisions.

   Interactive Brokers added ChatGPT (an AI assistant made by OpenAI) and Grok (an AI assistant made by xAI) to its platform. The move lets traders use these AI tools directly within their brokerage account to help with trading and investment research. Brokerages are embedding AI assistants into their platforms to help customers make faster decisions and stay competitive as AI tools become standard in financial services. This reflects a broader shift toward AI-powered tools in retail investing.

   The integration signals Interactive Brokers' strategy to offer multiple AI options to its users, giving them choice in which assistant to use for their trading workflow.

2. 2

   Salesforce introduces Agentic Advisor to defend its CRM business against rising competition from AI note-taking applications.

   Salesforce has launched Agentic Advisor, a new feature designed to help sales teams. The company is responding to pressure from AI-powered note-taking tools that are encroaching on the CRM market and threatening Salesforce's core business model. CRM platforms like Salesforce have traditionally been the central hub where sales teams log customer interactions and manage relationships. If note-taking AI tools can capture and organize this information independently, they may reduce the need for a dedicated CRM system, which directly threatens Salesforce's revenue and customer lock-in.

   The success of Agentic Advisor will determine whether Salesforce can retain its dominance in the CRM market or whether AI note-takers will fragment the sales workflow into separate tools. This shift reflects a broader challenge: traditional software platforms must integrate AI capabilities or risk being displaced by specialized, AI-first alternatives.

3. 3

   DeepSeek releases an open-weights reasoning model that ranks #2 globally, signaling that frontier AI capability is no longer locked behind closed commercial systems.

   DeepSeek released DeepSeek-R1, an open-weights reasoning model (AI that shows its working steps before answering) that achieved #2 ranking on the AIME 2024 benchmark with a score of 96 out of 100, and #2 on the MATH-500 benchmark. The model uses 32T–33T tokens of training data and requires 8.7× fewer FLOPs (a measure of computational work) compared with DeepSeek-V3.2 to operate. Open-weights models let any researcher or company download and run the AI themselves, rather than paying for access through an API or commercial service. A #2-ranked reasoning model in open form means the frontier of AI reasoning—a capability that was recently considered proprietary to well-funded labs—is now publicly available. This shifts the economics and accessibility of advanced AI.

   DeepSeek-R1 is released under an open license (MIT), meaning there are no licensing restrictions on use. The model's efficiency (using far fewer computational resources than its predecessor) may lower the barrier to deployment for organizations without massive cloud budgets.

4. 4

   AI agents are beginning to run in continuous self-improving loops, a shift Boris Cherny says is as significant as the move from hand-written code to AI-generated code—but the token costs could be steep.

   Claude Code creator Boris Cherny explained at Meta's @Scale conference that AI agents are now moving beyond writing code on demand to running in loops where they continuously prompt other agents to improve code architecture, unify duplicated abstractions, and submit pull requests—never stopping as long as compute is available. Agentic loops represent a shift from managing discrete AI tasks to authorizing swarms of agents to work endlessly in the background. The approach builds on existing techniques like recursive loops and the Ralph Loop (which asks an AI model if it has accomplished its goal), but applies non-deterministic logic where the AI decides when to stop rather than a clear programmed condition. However, because loops consume tokens much faster than chatbots and have no built-in spending ceiling, the costs may outweigh benefits unless token spend, drift, and other AI issues are carefully managed.

   Cherny framed loops as part of a broader trend in 'test-time compute'—the idea that models can solve nearly any problem if you throw enough computing power at them. For hill-climbing problems like code improvement, loops can keep making incremental advances until reaching a threshold or until compute runs out, making the financial sustainability of continuous agentic loops a key question for businesses beyond token-selling companies like Anthropic.

5. 5

   Google and Nokia are rolling out AI agents to help telecom operators manage networks faster, cutting problem-solving time by 50% to 80%.

   Alphabet's Google and Nokia expanded their partnership to integrate Google's Gemini AI models into Nokia's Assurance Center, a network management software platform. Nokia plans to deploy six specialized agents—two of which (Router and Event Triage) are already working—to help route network issues, triage events, select key performance indicators, detect anomalies, and build dashboards. The full platform is expected to launch as a software-as-a-service product on Google Cloud Marketplace in September 2026. Telecom networks generate massive amounts of data, and operators currently spend significant time sorting through alerts and troubleshooting. These AI agents address a real operational pain point by automating and accelerating problem diagnosis and resolution, which means lower costs and faster service recovery for network operators.

   The product launches on Google Cloud Marketplace in September 2026. For Google Cloud, this partnership demonstrates a new enterprise use case for Gemini beyond consumer chatbots; for Nokia, it strengthens its network software business with embedded AI capability.

6. 6

   Governing the Agentic Enterprise: What Marketing Leaders Need to Know Now

   Governing the Agentic Enterprise: What Marketing Leaders Need to Know Now