There are some simple tests to determine whether Amodei (and Altman, and Musk, and Zuckerberg, and Nadella, and Pichai) are serious about any commitment to safety, alignment, “pacing the frontier”, and so forth. Ask: How much are you, personally, willing to declare and accept responsibility for problems in your products, and the problems they cause for other people? Personally, not shielded by the protections and the pockets of the corporation? … Read more
This post originally appeared on LinkedIn. Language is the bottleneck. We need to be able to understand each other and ourselves if we want good things, or if we want other people to provide good things for us. Sometimes we need to express ourselves very precisely, effectively, and efficiently. That makes mathematical modeling the bottleneck. Mathematical notation can help us to make a lot of progress on that. Computers might … Read more
It might okay for an organization to use LLMs in technical support — to parse a request for content, and then provide links to curated answers, call services that can provide correct answers, or summon a human agent. Why do companies use GPTs — generative AI — for technical support? Remember Brooks on GPTs: a GPT is not designed to produce a correct answer; it’s designed to produce an answer … Read more
There are subtle but serious problems in the testing business right now, and if we don’t attend to them they will lead to bigger problems. The core problem I’m attending to today is confusion between testing and quality engineering. Engineering Let me start by saying I believe that quality engineering is a good thing. Or rather, I believe that engineering is a good thing. As Billy Vaughan Koen said in … Read more
“We’re happy to acknowledge when we make errors, and we’ve made a few in this case for sure, but also that we are willing to address those. And we’ve taken immediate action in all of these cases,” he said. “And we certainly apologize for any effect or adverse effect that this has had on trust in the community, because we want people to be confident in what we’re doing.” David … Read more
This open letter originally appeared as a post on LinkedIn. Dear QBD Books… Good software design matters. Without good software design for your Web site, users may have trouble, and that trouble can cost you money. Testing matters too. Testing experientially matters. Without experiential testing, you won’t discover certain things about your site that represent bad design. And that can cost you real money. Let me explain. Today, I wanted … Read more
Podcasts on Quality Assurance and Testing In this video podcast, Alex Khvastocvich sat down with James Bach and Michael Bolton to answer a deceptively simple question: What is testing? Alex says A lot of people still reduce testing to “finding bugs.” But if you take testing seriously, it’s much bigger than that. We talk about: What testing actually is (and what it isn’t) And a bit about the future of … Read more
Over the last couple of days, a LinkedIn conversation has reminded me to write about the Productivity Paradox, which appears in Taking Testing Seriously: The Rapid Software Testing Approach, by James Bach and me. (I haven’t written a blog post about the delivery and release of the book. I was on tour when it came out, and then a bunch of online classes, the holidays, some sad occasions in the … Read more
When my daughter was two years old, and when she saw something unusual, she would say, “Dis is weird!”, mispronouncing the word as “weeood”, as little kids are prone to do. Cute as a button. Unrelated to that, the other day I was thinking about what might happen if we had a body of text that we wanted to modify — say, in an article, or a requirements document, or … Read more
How might we use a GPT in testing? Some have suggested that we could use GPTs to generate test data. Randomized test data can help to reduce patterns of certain biases in our testing. We might assume that getting a bot to produce random data based on a straightforward prompt would be easy. So here’s a little one-off, first-hurdle experiment I performed July 24, 2025 using Google’s Gemini 2.5. Here, … Read more