An unreviewed preprint compares 14 language models with human writing and identifies six patterns that may make AI text harder to process.
Six Ways AI Writing Drains Readers, Plus How To Make It Easier
Browsing Category
AI & Work
1501 posts
Six Common AI Writing Patterns That Make Reading Harder
A non-peer-reviewed study compares 14 language models with human writing and identifies six patterns linked to harder-to-process sentences.
Can Open-Weight Models Make On-Prem AI Practical?
IBM says open-weight AI can approach frontier performance at lower token cost. Benchmarks, hardware expenses and staffing complicate the on-prem case.
Our Newest AI Quest: Mapping The Human Brain
Google Research and Stanford launch Polar Peaks, an AI Quest for ages 11–14 about brain mapping, data preparation and model testing.
Anthropic Eyes IPO After 14X Growth Surge
A New York Times headline reports 14-fold growth as Anthropic prepares for a possible IPO, but the metric, timeframe and listing status are unclear.
How To Read The CVE Rise Alongside AI And Patch Gaps
A sharp rise in 2026 CVE records is concentrated among major vendors. Attribution is limited, and reported vulnerabilities far outnumber known fixes.
Jack Clark Brings His Self-improving AI Forecast Forward By Several Years
An RSS headline says Anthropic co-founder Jack Clark has moved his self-improving AI forecast to 2027; the reasoning and details are unavailable.
Planning Your 2026 AI Workflow Automation Gear Setup
A 2026 gear checklist covers audio, video, storage, connectivity, input and power, with compatibility checks before any purchase.
AI And Human Oversight: 4 Questions For An Age Of Self-Improvement
A headline raises four questions about human oversight as AI systems improve themselves, but the available report provides no article body or supporting details.
A $345,000 Writing Job At Anthropic Focuses On Executive Communications
A CTech headline reports a $345,000 writing role supporting Anthropic executives, but the pay basis, duties and hiring status are unconfirmed.
