THE WORLD IS NOT STANDING STILL.RSS
BIG CHANGE.
BIG CHANGE, one tap away
BIG CHANGE, one tap away

Make room for the big picture.

Add BIG CHANGE to your home screen or desktop. It opens in its own window, with the latest stories when you’re online.

iPhone & iPad
Open bigchange.ai in Safari. Tap Share, then Add to Home Screen. Turn on Open as Web App if shown, then tap Add.
Android
Open bigchange.ai in Chrome. In the browser menu, choose Install app or Add to Home screen.
Desktop
In Chrome or Edge, use the install icon in the address bar or the browser’s app menu. In Safari on Mac, choose File → Add to Dock.
All briefings

Follow the change ·

The day,
in perspective.

What happened in AI and robotics — and the change behind it. A reading path through 50 stories from BIG CHANGE.

Ordered using RADAR attention, story opens and likes. Excerpts from our published reporting; open a story for the full context and sources. Dates in UTC.

guide

Evaluate an AI agent product across three decision axes

Map one agent product’s runtime and data paths, orchestration and oversight, and interface contract against your own mandatory requirements with a copyable decision worksheet.

The big change

What changed: Agent products offer several ways to run code, route work and connect tools. Their feature lists do not specify where a particular application runs or which permissions it grants. A three-axis map turns those choices into questions an owner can verify.

Why it matters: A documented handoff or approval feature has little value to a team whose deployed workflow does not use it. Connecting product evidence to the team's mandatory requirements exposes that gap before real data or consequential tool access is added.

What to watch: The decision changes only when a missing requirement is supplied, an unknown is resolved with evidence, or an authorized normal and failure trial verifies the configured workflow. A newly released version needs its own check.

Read the full story
—
Back to the briefing
guide

Create and check a daily news brief in Claude Cowork Projects

Save a personalized Cowork Project profile, check a current source-linked brief, then schedule one daily task with complete instructions and verify its first result.

The big change

What changed: Cowork Projects hold instructions and references, while Scheduled can run a saved task on a daily cadence. Claude documents both features.

Why it matters: A reusable profile can guide the first brief, and a complete copy of it can guide later runs. You still need to inspect real citations and the saved task.

What to watch: A scheduled entry does not prove that it used the intended context, found current reporting or delivered a checked result.

Read the full story
—
Back to the briefing
guide

How to dictate and review a Claude Code prompt before sending it

Prepare an authorized request, verify its captured words, and review the agent's result.

The big change

What changed: Hold mode offers a reviewable prompt before Enter.

Why it matters: You can compare the words on screen with a written task before giving the agent an instruction.

What to watch: A checked prompt is only an input. Review Claude's answer and any project changes separately.

Read the full story
—
Back to the briefing
analysis

Eko test: notes aided tower building

Eko note aided towers.

The big change

What changed: Eko's world and session reset hourly while its notes survived. Fixed weights rule out weight learning; erasing the recorded technique tests stored text's role.

Why it matters: A selected memory deletion reduced tower building in a designed simulator. The separate benchmark found that remembered advice could also lose to a matched retry, so memory itself cannot stand in for a score.

What to watch: A separate benchmark found verbal memory helpful in some computer-control settings and harmful in others. Comparing a remembered strategy with an equally budgeted fresh attempt tests whether the text adds value for the task at hand.

Read the full story
—
Back to the briefing
analysis

SpecEvo reserves costly model calls for program-search interventions

SpecEvo tests sparse frontier calls in program search.

The big change

What changed: Cheap proposals use Advisor guidance; Navigator redirects search.

Why it matters: Quality under a fixed call budget governs the decision. A team can compare allocations only if the same stopping rule and evaluator are used.

What to watch: The model bill leaves out the expense of running candidate programs. Faster proposal generation helps only if the evaluator can keep pace and its scores distinguish useful changes.

Read the full story
—
Back to the briefing
analysis

Google and Voltus: how flexible demand becomes capacity

The expanded PJM deal funds customer flexibility. Grid capacity depends on registration, accreditation and measured reductions.

The big change

What changed: Google is paying to assemble flexibility from other electricity customers as part of its response to data-center growth. That gives homes and businesses a role in supplying a service the company needs.

Why it matters: Demand response pays customers to reduce consumption when the grid needs relief. Payments reach them through a service provider, connecting their reductions to PJM and turning lower electricity consumption into a compensated grid service.

What to watch: Enrollment and performance can diverge. PJM's latest annual report records both registration shortfalls and failed tests despite strong aggregate test performance, making delivery evidence essential to judging a portfolio's contribution.

Read the full story
—
Back to the briefing
analysis

AI patent trial: seniors scored higher on redlining

Redlining tested patent lawyers' judgment.

The big change

What changed: Researchers measured later redlining separately from assisted drafting. Good tool-aided output did not establish unaided review judgment.

Why it matters: Claims set out the protection sought, while the application must describe the invention they cover. Reviewing an edit means checking that connection, beyond polished wording.

What to watch: Randomized endpoint comparisons can identify a group difference. An individual's growth and the durability of any difference require measurements at other times.

Read the full story
—
Back to the briefing
analysis

UserSWE tests coding agents in conversations

UserSWE rankings depend on turns.

The big change

What changed: A simulated user answers questions or reveals task facts after each agent turn, from a fixed pool. Feedback can redirect the repair.

Why it matters: Claude Opus 5 leads at five turns, GPT-6-Astra at eight; rewarding early passing favors Opus. Scoring changes the leader.

What to watch: Give candidates the same task and acceptance tests. Record when missing requirements were clarified and how much correction a passing repair needed.

Read the full story
—
Back to the briefing