A random mix from the whole library. Shuffle for a new set.
Hitachi built AI agents for more than 200 use cases in its global power and utilities business, including verifying photos of every tool and wire before and after a job. Frontline worker productivity rose 400 percent and safety improved.
A man with no coding experience used AI to generate a VBA script in Excel. The script fixed a manual data entry gap at his work and took the risk of typos and misreads out of the job.
The AI enabled a camera on a test device and he taught it to reboot into the previous kernel on build failure so it could iterate for days; ended with a usable setup and camera login working.
Asana says Piper, its Agentforce agent, drives four times the sales conversation volume and took about 45 days to deploy, per Salesforce's release.
A mathematician with more than a dozen finished but never submitted papers pulled out four of them and shared them with AI. The AI caught errors, built a spreadsheet comparing his new concepts against related papers, converted plain TeX to LaTeX, wrote abstracts he adopted almost word for word, and fixed bibliography entries unprompted. All four papers improved to the point where he submitted them.
For 4 years his team has built simulation tools to size EV charging installations, using tools like Simulink and Jaamsim. Yesterday he built it in web format in minutes using AI.
The AI runs his entire video production pipeline via an AI plugin: full 5-minute explainer videos from one prompt, thumbnails and titles, auto-upload via the YouTube API.
He built Stackboard, a free Chrome new tab page that shows your own bookmarks as a board. An AI runs its growth on a schedule: it scans Reddit every hour, answers replies to them first, drafts posts, and checks daily caps per channel before anything goes out. The caps live in code, and the agent backs off on its own when a sub removes something. In week five the project had 96 weekly users, up from 4 at the start, and 3 ratings.
Asked the AI to break his team's own Postgres MCP guards and it found pg_logical_emit_message, which writes to the WAL inside a read-only transaction, leading to a hardened replacement.
His free French teaching site Ralentir travaux, used by millions of visitors, had been more or less abandoned because its old code ran to thousands of pages. Over about six months the AI and the AI cleaned, fixed and modernised the code and added new tools, and the rebuilt site relaunched for its 20th anniversary.