New model
IMPACT 58Long-running agentic coding became materially more reliable
Claude Opus 4.1 extended sustained autonomous performance on multi-step coding and research tasks, with providers reporting multi-hour task horizons.
What is happening right now
Every event below is verified, classified and scored by our internal analysis pipeline, then mapped to the careers it affects. No raw feeds, no unvetted links - just the verdict.
Feed
Disruption feed (30 events)
Our verdicts, newest first - with disruption scores and affected careers. Underlying research and data collection methods stay proprietary.
- New modelIMPACT 56
Claude Opus 4.1 — Anthropic
Multi-hour autonomous coding and research tasks with high task-completion rates.
- Labour market update
Annual employment and wage data by occupation updated
The OES programme published its latest annual employment and median wage estimates for detailed SOC occupations.
- New modelIMPACT 55
ElevenLabs v3 (alpha) — ElevenLabs
Commercial-quality synthetic narration and voice acting with multi-speaker dialogue.
- New modelIMPACT 55
Claude Sonnet 4 — Anthropic
Sustained multi-step agent tasks, large codebase editing and tool chaining.
- New modelIMPACT 53
Veo 3 — Google DeepMind
Generates short video clips with matching speech and sound effects from text prompts.
- New modelIMPACT 48
Video generation with synchronised audio and dialogue
Veo 3 generates video clips with matching speech and sound, removing a separate audio post-production step for short-form content.
- Research publishedIMPACT 20
Detecting and countering malicious uses of Claude
Detecting and Countering Malicious Uses of Claude
- New modelIMPACT 56
o3 — OpenAI
Extended chain-of-thought reasoning for maths, science and code; tool use during reasoning.
- New modelIMPACT 55
GPT-4.1 — OpenAI
Large-context document work, improved coding performance and instruction following.
- New modelIMPACT 52
Llama 4 Scout & Maverick — Meta
Open-weight multimodal generation at low inference cost.
- AI adoption signalIMPACT 54
Business adoption of AI rose again in 2025 surveys
The AI Index reports continued increases in the share of organisations using generative AI in at least one business function, alongside rapid benchmark gains on reasoning and coding tasks.
- New modelIMPACT 55
Gemini 2.5 Pro — Google DeepMind
Video and audio understanding, long-context reasoning and strong maths/code results.
- Labour market updateIMPACT 66
After image-generation adoption, fewer active graphic designers and lower earnings
Studying a large stock-content platform around the integration of image-generation models, the authors find a decline in the number of active graphic designers and in their earnings, with effects concentrated in routine and less-distinctive work.
- AI adoption signalIMPACT 68
Measured AI usage concentrated in software development, writing and analysis
Aggregated and anonymised Claude conversation data mapped to O*NET tasks shows that usage is concentrated in a small number of occupations and tasks — notably software development, writing and technical analysis — and that a substantial share of usage is augmentation rather than full automation of a task.
- New AI toolIMPACT 46
Claude Code — Anthropic
Autonomous multi-step implementation and refactoring in real repositories.
- AI adoption signalIMPACT 76
AI assistant raised support-agent productivity 14%, with the largest gains for novices
In a randomised controlled trial with about 5,000 customer support agents, access to a generative AI assistant increased issues resolved per hour by about 14% on average, with gains of roughly 34% for the least experienced workers and negligible effects for the most experienced.
- New modelIMPACT 54
DeepSeek-R1 — DeepSeek
Openly licensed reasoning capability, previously restricted to closed APIs.
- New AI toolIMPACT 41
Zapier Agents — Zapier
Natural-language orchestration of cross-application business workflows.
- New modelIMPACT 52
DeepSeek-V3 — DeepSeek
Near-frontier text and code capability at a fraction of typical serving cost, changing the unit economics of automation.
- New modelIMPACT 50
Sora — OpenAI
Short-form video generation from prompts and stills, with editing controls.
- New modelIMPACT 52
Text-to-video generation became commercially available
Sora made short text-to-video generation publicly available, followed by competing systems with synchronised audio in 2025.
- Research publishedIMPACT 74
Experienced open-source developers completed 26% more tasks with AI
In pre-registered experiments with experienced open-source developers, access to AI tools increased the number of tasks completed by about 26% relative to control.
- New modelIMPACT 50
Suno v4 — Suno
Library-quality music beds and full songs generated from a prompt.
- New modelIMPACT 53
Stable Diffusion 3.5 Large — Stability AI
Self-hostable production image generation, inpainting and style control.
- New AI toolIMPACT 43
Replit Agent — Replit
End-to-end app scaffolding, deployment and iteration without local tooling.
- New modelIMPACT 50
Qwen2.5 — Alibaba
Low-cost multilingual generation and translation across dozens of languages.
- Labour market updateIMPACT 58
Data entry occupations are already among the fastest-declining projections
BLS ten-year employment projections list data entry keyers and closely related clerical occupations among the occupations with the largest projected employment declines.
- New capabilityIMPACT 72
Open-weight image models reached production-quality text rendering
FLUX.1 and subsequent open-weight diffusion models produce legible text and layout-consistent imagery, making them usable for marketing collateral rather than only concept art.
- New modelIMPACT 54
FLUX.1 — Black Forest Labs
High-fidelity image generation with strong text rendering, usable for layout and brand mockups.
The Live Disruption Radar streams continuously. Each entry shows our independent assessment - impact score, confidence and affected careers.