AI & Technical question

How would you make the internet faster?

Practice this question out loud. An AI interviewer asks it, follows up like a real interviewer would, and scores your answer. Type or speak.

What this question tests

Tests technical breadth on internet infrastructure, requiring the candidate to identify real bottlenecks rather than a vague make it faster answer.

Clarify scope: last-mile access speed for end users, or backbone and routing latency, since the fix differs.
Identify the main bottlenecks: last-mile bandwidth in underserved areas, DNS resolution time, and content distance from origin servers.
Segment by geography: urban areas with fiber already available versus rural areas relying on satellite or older infrastructure.
Propose layered fixes: expanding CDN edge locations to cut content distance, promoting protocols like HTTP/3 and QUIC to reduce connection overhead, and investing in last-mile fiber or low-earth-orbit satellite for underserved regions.
Weigh cost and who pays: ISPs, governments, or platforms subsidizing infrastructure in low-margin regions.
Define success as reduced median page load time and latency, segmented by region.

Breaks a vague prompt into concrete technical levers, CDN placement, protocol efficiency, and last-mile access, instead of one blanket answer.
Distinguishes latency problems, fixable with protocol and routing changes, from bandwidth problems, fixable only with physical infrastructure.
Uses illustrative numbers as assumptions: assume rural median latency is 3x urban, so infrastructure investment matters more there than protocol tweaks.
Names real technologies, HTTP/3, QUIC, and edge CDNs, showing actual technical grounding.

Giving a vague answer like build more infrastructure without naming specific levers.
Ignoring that bandwidth and latency are different problems with different fixes.
Not addressing who funds infrastructure investment in low-margin, underserved areas.

Chapters of the AI PM course that teach what this question tests.