Mike Taylor
PulseAugur coverage of Mike Taylor — every cluster mentioning Mike Taylor across labs, papers, and developer communities, ranked by signal.
7 day(s) with sentiment data
-
Anthropic releases Fable 5.2 with speed and capability upgrades
Anthropic has released its Fable 5.2 model, designed for large, long-running tasks that require minimal oversight. Early users report significant speed improvements and fewer content blocks compared to previous versions…
-
Every hosts Thesis: 2027 conference on AI's future impact
Every, a platform focused on AI and its implications, hosted its inaugural conference, Thesis: 2027, featuring discussions on the future of work and business in an era of increasing automation. The event brought togethe…
-
AI management training: Prompting principles enhance human team leadership
Working with AI agents is providing a practical, consequence-free training ground for management skills. The principles of effective AI prompting, such as giving clear direction and specifying output formats, are direct…
-
Every publication explores AI agents for content recommendation and task management
Every, a publication, is exploring the use of AI agents in various capacities. Their team is testing how to get agents to recommend articles, with one method involving rewriting titles and subtitles to directly answer p…
-
Anthropic's revenue model hinges on AI models younger than six months · 1 source tracked
Anthropic's S-1 filing reveals a significant reliance on its newest AI models, with those older than six months generating less than 20 percent of its revenue. This indicates a strategy where customers are expected to c…
-
Every explores personal vs. shared AI agents and performance evaluation
This newsletter discusses two approaches to using AI agents within a company: individual agents for personal use and a shared agent for the entire team. It highlights methods for evaluating agent performance, such as de…
-
Every establishes AI frontier team to manage rapid advancements
The company Every has established a dedicated "frontier team" to manage the rapid advancements in AI, allowing the rest of the company to maintain focus on their core responsibilities. This team experiments with new AI …
-
The Every launches shared AI coworker for company-wide Slack integration
The Every launched its new AI coworker, the Every Agent, designed to be shared across an entire company and operate within Slack. This agent allows teams to assign tasks and save successful workflows as reusable skills.…
-
Author details shift to AI agent orchestration for complex tasks
The author describes a shift in their AI usage from simple prompting to more complex agent orchestration, inspired by a framework for AI adoption. Initially, tasks were handled in single chat threads, but now multiple s…
-
Every CEO encourages ambitious AI experiments despite high token costs
Every CEO Dan Shipper is encouraging employees to experiment more freely with AI, even if it leads to higher token consumption. He believes that exploring new AI capabilities and workflows could unlock significant produ…
-
Astra AI model gains traction with advanced computer use capabilities
Astra, an AI model capable of operating applications, has shifted from being a yellow-rated tool to a daily driver for Mike Taylor, head of evaluations at Every. This change was prompted by Astra's enhanced 'computer us…
-
Every builds personalized AI benchmarks for employee tasks
Every, a company focused on AI evaluation, is developing personalized benchmarks for its employees to assess AI model performance on specific job tasks. CEO Dan Shipper explained that these custom evaluations, designed …
-
Every explores AI-assisted writing with new 'Compound Writing' plugin
The publication Every is exploring how individuals write with AI, framing it as a collaborative process rather than a question of AI authorship. They are launching a series called "How We Write Now," beginning with a fo…
-
AI adoption challenges for executives detailed in new Every guide
Every, a consulting practice focused on AI adoption, has compiled a document answering 33 common questions executives have about implementing AI. These questions, gathered from interactions with over 100 companies inclu…
-
Every builds AI clones of employees to automate expertise
Every, a media company, is developing AI agents to replicate the expertise of its key personnel. Initially, they created KateBench to emulate editor Kate Lee's copyediting skills using GPT-5.6 and historical data. The c…
-
New AI Trick Exposes Model Reasoning, Potential Data Leaks
Researchers have developed a method to extract internal reasoning processes from AI models, revealing potential vulnerabilities. This technique can expose personal information, such as passwords and API keys, though thi…
-
Anthropic's Fable model excels at uncovering hidden task assumptions
The AI model Fable, from Anthropic, is being highlighted for its ability to uncover hidden decisions and assumptions within complex tasks. Unlike models that excel at executing predefined plans, Fable is particularly ad…
-
Every details Codex use cases as Anthropic's Fable 5 returns
The AI company Every is highlighting practical use cases for its Codex agentic workspace, emphasizing its versatility and ease of use for non-technical users. Head of consulting Natalia Quintero shared how Codex has bee…
-
US restricts GPT-5.6 Sol access; AI agents expand to knowledge work
OpenAI has released GPT-5.6 Sol, but access is restricted by the U.S. government to about 20 pre-approved companies. This rationing is expected to extend to token access, which will be allocated based on demonstrated re…
-
AI enters allocation era: Token budgets to be managed like trading portfolios
Companies are shifting from aggressive AI adoption to a more strategic allocation of resources, as the cost of powerful models and long-running agents leads to significant enterprise bills without tangible results. This…