Topic: token efficiency
-
Google unveils Gemini 3.6 Flash and security AI, teases 3.5 Pro and Gemini 4
Google released Gemini 3.6 Flash as a refined replacement for the deprecated 3.5 Flash, offering marginal improvements in coding and multimodal performance based on user feedback, while the long-awaited Gemini 3.5 Pro remains delayed. The new model achieves higher benchmark scores (49% on DeepSWE...
Read More » -
From AI Discovery to Agentic Commerce: Winning the Decision Layer
AI-referred traffic to retail sites surged 4,700% year-over-year by mid-2025, and AI influenced one in five online orders during Cyber Week, driving $67 billion in sales. Brands must optimize for the "AI decision layer" by ensuring machine accessibility, token efficiency (e.g., llms.txt files), a...
Read More » -
Microsoft launches Web IQ: Bing-powered search for AI agents
Microsoft launched Web IQ, a new grounding API powered by Bing's index that is rebuilt from the ground up for speed and efficiency, providing AI agents with fresh, real-time data from the web. Unlike traditional search APIs designed for humans, Web IQ is optimized for AI agents, prioritizing effi...
Read More » -
Microsoft Bing Grounding APIs Now Power AI Agents
Microsoft launched Web IQ, a set of grounding APIs that allow AI agents to directly access Bing’s search index for real-time, factual data, delivering passages and structured evidence instead of full web pages to improve efficiency and reduce costs. The system achieves sub-165 millisecond respons...
Read More » -
Glean’s revenue hits $300M as AI cost-cutting drives growth
Glean has reached $300 million in annual recurring revenue, tripling its $100 million milestone from 15 months ago, despite increasing competition from tech giants like Google, Microsoft, and OpenAI. The company differentiates itself through a "context graph" that deeply understands each customer...
Read More » -
Chrome Lighthouse now checks for llms.txt
Google's Lighthouse update introduces an "Agentic Browsing" audit category that checks for llms.txt files to improve discoverability and efficiency for AI agents, distinct from traditional SEO crawling. While Google advises that llms.txt is unnecessary for search rankings, Chrome's own audits now...
Read More » -
Why Google Uses Markdown for Developer Docs, per Mueller
Google uses markdown pages for developer documentation to help AI coding systems parse reference material more efficiently, not as an SEO strategy. For non-developer sites, Mueller advises against creating markdown versions, stating they won't drive sales and are only useful for competitors. Muel...
Read More » -
Claude Code product lead on usage limits, transparency, and lean harness
Richard Sutton's "The Bitter Lesson" essay, which argues that general-purpose AI approaches scaling with compute outperform domain-specific methods, serves as a guiding principle for the team. The team aims to quickly move from user conviction to product shifts within a week, and envisions Claude...
Read More » -
OpenAI GPT-5.5 vs Claude Opus 4.7: Full Comparison
GPT-5.5 leads on most standard benchmarks like Arc Prize, Terminal-Bench 2.0, and Humanity's Last Exam, while Claude Opus 4.7 excels specifically in agentic coding tasks, scoring higher on SWE-Bench Pro. GPT-5.5 offers broader features through ChatGPT, including image generation via ChatGPT Image...
Read More » -
OpenAI’s GPT-5.5 Boosts Efficiency and Coding Performance
OpenAI launched GPT-5.5, described as its most intuitive model yet, capable of handling complex multi-step tasks like coding, research, and cross-tool work with improved safeguards and efficiency. The rollout begins Thursday for Plus, Pro, Business, and Enterprise ChatGPT tiers, with a more power...
Read More » -
Betterleaks: Open-Source Secrets Scanner for Enhanced Security
The creator of the popular Gitleaks tool has launched a new open-source secrets scanner called Betterleaks, designed as a direct, compatible replacement for detecting exposed credentials in code. Betterleaks introduces a novel filtering technique called Token Efficiency, which uses byte pair enco...
Read More » -
OpenAI Launches GPT-5.4: Supercharged for Knowledge Work
OpenAI has launched GPT-5.4, a major update featuring specialized variants like GPT-5.4 Thinking and GPT-5.4 Pro to address complex tasks and compete in a crowded AI market. A key advancement is enabling **agentic workflows** for computer interaction, allowing the model to automate digital tasks ...
Read More » -
Vectorization & Transformers: The Core of Modern Information Retrieval
Modern search engines have evolved from keyword matching to interpreting user intent and concepts, primarily through semantic understanding powered by machine learning and models like the vector space model. Core technologies enabling this include TF-IDF, cosine similarity, and transformer archit...
Read More » -
Cloudflare's AI Agents Spark SEO Concerns
Cloudflare's new "Markdown for Agents" feature automatically converts web pages into a simplified markdown format for AI crawlers, aiming to reduce processing costs and improve efficiency for machine readers. The feature has raised significant SEO concerns, as it could enable "AI cloaking" by all...
Read More » -
Cloudflare's AI Bot Markdown: A Complete Guide
Cloudflare's new "Markdown for Agents" feature automatically serves a lightweight markdown version of web pages to AI crawlers upon request, reducing the computational tokens needed for processing without requiring separate pages from site owners. The service uses standard HTTP content negotiatio...
Read More » -
We Tracked 10 Sites: Does llms.txt Matter?
AI crawlers from major providers like Google and OpenAI rarely request llms.txt files, and no leading LLM company has officially committed to using the standard for content discovery. A 90-day study found that implementing llms.txt had no measurable impact on AI traffic for most sites; any observ...
Read More » -
Google Gemini 3 Flash Now Powers Default App & AI Mode
Google has launched Gemini 3 Flash as the new default AI model for all free users in the Gemini app and for AI Mode in Google Search, making advanced AI more accessible. The model is available for developers via the Gemini API at a cost of $0.50 per million input tokens and $3.00 per million outp...
Read More » -
OpenAI's Codex Max: Faster AI Coding, Fewer Annoyances
OpenAI has launched Codex Max, an upgraded AI coding model that offers faster execution, reduced token use, and better handling of complex tasks, available to various subscription tiers. The model features compaction technology to manage much larger workloads by intelligently compressing context,...
Read More » -
Boost Your Coding Speed & Save with GPT-5.1
OpenAI's GPT-5.1 enhances coding efficiency and reduces costs through smarter reasoning modes and extended prompt caching, addressing latency and expense issues for developers. The update introduces adaptive reasoning and no reasoning mode, which adjust cognitive effort based on query complexity ...
Read More »