Top stories

GPT-5.3-Codex-Spark is a new model optimized for real-time coding, capable of processing over 1000 tokens per second on ultra-low latency hardware. This model is a smaller version of GPT-5.3-Codex and represents the first milestone in a partnership with Cerebras.

AI currently outperforms humans in narrow domains and acts as a significant amplifier of human creativity and productivity. However, high latency and cost impede its widespread adoption, with language models lagging behind human cognitive speed and coding assistants causing disruptions in workflow.

Lars Faye argues that AI coding assistants can weaken the skill formation novice developers need to use those tools safely and effectively. He calls this the “expert novice” problem: developers entering the field alongside large language models are urged to use AI to keep pace, while effective prompting, code review, system design and output verification still require experience developed through repeated problem-solving. Faye says experienced engineers currently gain more from the tools because they can steer and audit their outputs. Faye cites the study “The Widening Gap: The Benefits and Harms of Generative AI for Novice Programmers,” highlighted by JetBrains, which found that participants using heavier AI assistance often skipped planning, developed an “illusion of competence,” and became lost in generated solutions. Participants who limited assistance performed better by rejecting unhelpful suggestions and using AI to accelerate solutions they already understood. He also cites a 2025 University of Pennsylvania study of 1,000 mathematics students, which reported that unrestricted LLM use led to test performance 17% below a textbook-only group, while a tutor-oriented GPT condition improved AI-assisted practice results by 127% but produced test scores similar to the textbook group. Faye recommends using LLMs primarily for interactive documentation, tutorials and Socratic exercises rather than routine code generation when learning. He says developers should verify AI outputs through official documentation, peers and hands-on testing, and distinguish delegating tedious work from delegating judgment.

Git-memento is a Git extension that records AI coding sessions associated with commits. It stores a cleaned markdown conversation as a git note on each new commit, allowing for human-readable documentation of the coding process.

Claude Code is gaining popularity within Microsoft for its user-friendly interface, appealing to both developers and nontechnical users. Comparisons with tools like Anysphere’s Cursor and GitHub Copilot show that while no single AI coding tool excels in all areas, Claude Code is emerging as a preferred option.

AI coding at home can be done affordably by self-hosting, which involves purchasing hardware and running open-source models locally without ongoing costs. This method requires a significant upfront investment and may offer less powerful models compared to those from leading labs.

AI coding tools can quickly generate visually appealing code but struggle with detailed implementation. Extended use of AI coding agents has led to a substantial portfolio of projects, showcasing their potential and limitations.

AI coding tools significantly enhance developer productivity but come with hidden costs. Developers currently operate between manual coding and fully autonomous AI systems, with the balance shifting towards greater AI integration.

Spotify's top developers have not written any code since December, attributing this shift to the use of AI in their development processes. The company has released over 50 new features and changes to its streaming app during this period.

Anish Acharya, a partner at A16z, argues that using AI-assisted coding for all business functions is not worthwhile and suggests that AI should concentrate on core business development rather than rebuilding enterprise software. He also noted that recently oversold software stocks are likely to recover.