doorss

GitLab announced workforce reductions and structural changes tied to the agentic era, planning to rebuild around AI-native workflows and reduce headcount accordingly.

Amazon employees are reportedly using the in-house AI tool MeshClaw for unnecessary tasks to climb the company's internal AI usage leaderboard.

James Shore argues that AI coding agents must reduce maintenance costs, not just speed up code writing—if you write code twice as fast, you'd better have halved your maintenance burden too.

A New York Times report describes how Meta's aggressive push to integrate AI into every workflow is causing widespread dissatisfaction among employees.

The FT argues that whether AI can perform a task is only a small part of the employment picture, and the 'jobpocalypse' narrative overlooks many practical barriers to job displacement.

Jason Koebler writes about how AI-generated writing online is becoming inescapable, mentally exhausting to filter, and is starting to distort regular human communication.

A research study examining the measurable ways LLMs are changing written language patterns across the web.

A post examining how AI sycophancy persists despite surface-level improvements, referencing a Stanford paper on how sycophantic AI advice leads to worse decisions.

France is advancing legislation that would undermine end-to-end encryption in messaging apps.

Meta has removed end-to-end encryption from Instagram direct messages.

The EU Parliamentary Research Service has labeled VPNs as a loophole to be closed in the context of age verification efforts.

Google Chrome has been found to silently download and install a 4 GB Gemini Nano AI model on users' devices without explicit consent.

Anthropic research on providing Claude with explanations for its training objectives, exploring how understanding the 'why' behind guidelines affects model behavior.

Anthropic published research on natural language autoencoders that convert Claude's internal representations into readable text, advancing interpretability.

The Anthropic Institute (TAI) shares its research agenda for investigating AI's real-world impact, using information from within a frontier lab to share findings publicly.

A survey of 81,000 Claude users finds that people in AI-exposed roles have greater concerns about job displacement, with concerns higher in lower-income countries.

Bluesky published technical details on how they serve the 'For You' algorithmic feed, one of the platform's distinguishing features where anyone can run their own custom feed.

A detailed breakdown of how ChatGPT's advertising system works, including the full attribution loop from ad placement to conversion tracking.

TV advertising is becoming increasingly valuable as streaming platforms like Netflix and Disney+ adopt ad-supported tiers and look for new revenue sources.

Mielipidetutkimuksen mukaan suomalaiset eivät luota Yhdysvaltoihin ja haluaisivat hankkia aseita muualta. Grönlanti-asiassa suomalaiset tukisivat Tanskaa.

Etla arvioi Iranin sodan taloudellisen vaikutuksen Suomeen pieneksi ja ennustaa kasvua teollisuuden lisäksi myös rakentamiseen ja palvelualoille.

THL:n selvitys näyttää, miten kännykkäkielto on parantanut koulun ilmapiiriä: korttipelit korvasivat puhelimet ja opiskelu välitunneilla lisääntyi.