Sector
Work and software
Software is the first kind of work where agents are measured on real tasks at scale. It shows early what will happen to other kinds of work.
Our predictions
Our view, with how sure we are. Each one links to the evidence it rests on.
On the road to AGI
AI agents will reliably finish software and office tasks that take a skilled person a full working day.
What changes
Most routine software maintenance, such as fixing bugs and updating code, will be done by agents and reviewed by people.
What changes
Demand for early-career workers in the most AI-exposed jobs will keep falling, while experienced workers in the same jobs hold steady.
On the road to AGI
Open standards for connecting agents to tools and to each other will become as common for agents as web standards are for websites.
On the road to AGI
Checking work, not producing it, will become the main limit on how much we can trust autonomous AI.
Progress around the world
Public results from other teams that shape our view.
· METR
METR's frontier risk report: agents saturate the long-task test, and some cheat
The best public models reached a 50% time horizon of about 12 hours, but only about 1.5 hours at 80% reliability. At least 16% of successful runs on 8-hour-plus tasks were cheating on review. METR judged a direct AI takeover implausible for these systems.
What it means for AGI: agents can now attempt day-long work, but reliability and honest checking lag far behind. Measuring progress is getting harder as tests saturate.
Source: metr.org(opens external site)
· Stanford Institute for Human-Centered AI
AI Index 2026: capabilities jump, but unevenly
Agent success on Terminal-Bench rose from 20% to 77.3% in a year, and agents solved 93% of cybersecurity tasks, yet robots succeeded at only 12% of real household tasks. AI data-centre power capacity reached 29.6 GW, and four in five US students use AI for schoolwork.
What it means for AGI: digital abilities are racing ahead while physical abilities and governance lag.
· METR
METR updates its time-horizon measure: doubling every 3 to 4 months since 2024
With a larger task suite (228 tasks), METR estimates the length of task agents finish with 50% success doubled about every 89 days since 2024, faster than the roughly 7-month long-run trend.
What it means for AGI: progress on long tasks sped up, not slowed down, through 2025.
Source: metr.org(opens external site)
· Linux Foundation
Open standards for AI agents get a neutral home
The Linux Foundation formed the Agentic AI Foundation around the Model Context Protocol, goose and AGENTS.md, backed by major AI and cloud companies.
What it means for AGI: agents from different makers are getting shared ways to connect to tools and to each other, a precondition for large agent societies.
· Stanford Digital Economy Lab (Brynjolfsson, Chandar and Chen)
Early-career jobs fall in the most AI-exposed occupations
Using payroll data, the authors find a relative employment decline for workers aged 22 to 25 in the most AI-exposed jobs, while older workers in the same jobs held steady or grew. They call it early evidence.
What it means for AGI: the first effects on work are showing up at the entry level.
· Google DeepMind
AlphaEvolve: a coding agent that discovers better algorithms
An evolutionary agent built on Gemini models found a way to multiply 4×4 complex matrices with 48 multiplications, improving on Strassen's 1969 result, and improved on known answers for a share of over 50 open maths problems.
What it means for AGI: a loop of models proposing, testing and keeping the best ideas can produce new knowledge, an early form of self-improvement.
· METR
The length of tasks AI can finish is doubling about every 7 months
METR measured how long a task (in human working time) AI agents can complete with 50% success, and found it roughly doubled every 7 months over six years.
What it means for AGI: if the trend holds, agents move from minute-long tasks to day- and week-long projects. Whether it holds is one of the most important open questions.
Source: metr.org(opens external site)
· Princeton University and University of Chicago researchers
SWE-bench: AI agents measured on real software issues
A benchmark built from real GitHub issues in open-source Python projects, used to test whether AI agents can fix real bugs end to end.
What it means for AGI: agents are now judged on real work, not quiz questions. Scores have risen fast since it was released.
· Microsoft Research and collaborators (arXiv)
AutoGen: multi-agent conversation for complex tasks
A framework where several agents with different roles talk to each other, use tools and involve people to solve tasks.
What it means for AGI: building with groups of agents has become a mainstream engineering pattern.
Source: arxiv.org(opens external site)
· Google DeepMind (published in Nature)
AI discovers faster sorting algorithms used in a standard C++ library
A reinforcement-learning agent found sorting routines faster than human-written ones for short sequences; they were added to the LLVM libc++ library.
What it means for AGI: AI can improve the basic tools computing is built on, a small step toward systems that improve themselves.
Source: nature.com(opens external site)
· Princeton University and Google researchers (arXiv)
ReAct: language models that reason and act with tools
Showed that letting a model alternate between reasoning steps and tool actions improves results on question answering and decision tasks.
What it means for AGI: a model plus tools in a loop is more capable than a model alone. Most agent systems today build on this idea.
Source: arxiv.org(opens external site)
Publications
No publications yet.
Collaborations
Open to collaboration.