A crisis of expectations is brewing at the tech giant Google. The company has been forced to admit that the deployment of its flagship model, Gemini 3.5 Pro, is lagging behind schedule by several months. This decision resulted from an internal assessment: engineers needed more time to refine the system, particularly in the critical segment of code generation.
Panic Amidst Competitors' Success
The delay is causing serious concern within the corporation. According to Bloomberg, Google employees, researchers, and top executives fear that the company risks losing its strategic advantage in the market. While Google refines its product, major competitors—OpenAI and Anthropic—are actively releasing models that already surpass the search giant's current offerings in terms of capabilities.
The situation is exacerbated by the fact that OpenAI and Meta Platforms have already unveiled new solutions demonstrating higher efficiency in programming tasks. Google's attempt to rectify the situation by updating training data at the end of last month reportedly yielded disappointing results, according to informed sources. The market reacted instantly: the company's stock price dropped by 3.2%.
Internal Barriers and Regulatory Pressure
One of the reasons for the missed deadlines is the complex corporate structure. Numerous divisions are involved in preparing new models, and integrating AI into a wide portfolio of products—from the search engine to YouTube—creates logistical delays. Former employees describe the attempt to synchronize the work of different teams as trying to "boil the ocean".
Aside from internal disagreements, Google faces external pressure. The company is negotiating with the US government, which has tightened control over advanced AI models. Earlier this year, Anthropic and OpenAI already faced the need to restrict access to their products due to national security concerns and risks to IT infrastructure vulnerabilities. Google confirms that it is testing the 3.5 Pro and Flash models with partners and engaging with regulators.
The Battle for Code: Humans vs. Algorithms
A genuine battle of approaches has unfolded within Google. On one hand, management is striving to accelerate the implementation of AI for code generation. On the other, there are competing teams within the Google Cloud, DeepMind, and Android divisions that are duplicating each other's efforts. Furthermore, part of the engineering community opposes total automation, insisting that core code must be written by humans to maintain quality standards.
Previously, employees faced restrictions on using Gemini for code analysis due to fears of data leakage into training sets. Although the policy has been relaxed, these restrictions significantly slowed down the experimental phase of development. Nevertheless, the company states that 75% of software code is now generated with the help of AI and continues to work on launching updates.
