How we work
Editorial independence
Why our numbers and opinions look different from the vendor blogs — and the explicit rules we hold ourselves to so they stay that way.
The structural reality
Every vendor in quantum computing — IBM, Google, Microsoft, Quantinuum, IonQ, Xanadu, Rigetti, QuEra, Pasqal — has a primary product to sell: access to their hardware. Their educational content, however well written, cannot publish three things:
- Side-by-side comparisons that show a competitor winning on any metric.
- Negative results — where their flagship paradigm (e.g., quantum machine learning) doesn't beat a classical baseline.
- Honest assessments of timelines that contradict their roadmap.
Those three categories happen to be the most useful information for a developer choosing a tool. That's the gap we exist to fill.
What we don't do
- No vendor exclusivity. No one gets a "preferred partner" badge. If we cover hardware, we cover all of it.
- No paid placements masquerading as editorial. Sponsored content is labeled, capped at one slot per email or page, and still has to clear the same technical bar as anything else.
- No advance review of editorial content by sponsors. If a sponsor disagrees with how we covered them, they can buy a paid slot to respond. They don't get to change the article.
- No undisclosed affiliate links.
We don't run any today. If we ever do, every affiliate link will carry
a visible
(affiliate)tag. - No NDA-gated coverage. We won't publish anything we can't independently link to a primary source.
What we do
- Cite primary sources. Vendor
calibration pages, arXiv papers, NIST publications, IETF drafts.
Press-release-only claims get marked
[unverified]. - Run our own code. Every benchmark on the site comes from a notebook we ran ourselves, on hardware we named. No "according to vendor X" without independent verification.
- Take positions. We pick the side we think is right. If you read tutorial 17 (QML vs XGBoost), tutorial 20 (hardware comparison), or the Hardware Honesty Tracker, you'll see opinions, not committee-edited fence-sitting.
- Update when we're wrong. Errors get fixed in-line with a dated correction note, never silently. The Edit-on-GitHub link at the bottom of every tutorial is real.
Sponsorship policy
We accept sponsorship from companies whose products we'd recommend anyway. The current best-fit list is on the sponsors page.
Sponsorship buys placement, not influence. Sponsored content is labeled "Sponsored" at the top, capped at one slot per newsletter or per page, and must pass editorial — meaning we will, on occasion, decline sponsorship money rather than publish something that doesn't help readers.
Conflicts to disclose
As of April 2026:
- No paid sponsorships active.
- No equity in any quantum-computing company.
- No consulting engagements with hardware vendors.
- No employment with any vendor named on this site.
- Free-tier accounts on IBM Quantum (the 10-min/month tier — same as any reader has access to).
This list will change as the site grows. When it does, this page gets updated, every relevant article gets a disclosure footnote, and the change is announced in the next newsletter.
Why this matters to you
If a vendor blog tells you "QML matches classical performance" and we tell you "XGBoost wins by 5 points on Wisconsin Breast Cancer," only one of those claims comes with running code, a verifiable benchmark, and no financial stake in the answer. We work hard to be the source you trust when accuracy matters more than vendor relations.
Questions or concerns about a specific piece of coverage? [email protected]