The author's nearly incomprehensible complaint (http://blog.armorize.com/2010/08/iframes-and-url-stringency-mozilla.html) is essentially that this is allowed to load, while entering http://foo:firstname.lastname@example.org in the address bar results in a phishing-related warning. The purpose of this warning is to confirm you actually understand the syntax of the URL displayed in that very address bar.
Let that sink in for a while.
If you don't see a fundamental difference between these cases that makes this report completely rubbish, you should probably surrender your geek badge now.
I have a fairly long track record in the security industry, and I'm really puzzled by Cringely's assertion. It's hard to tell if he is trying to make a point out of a semantic squabble, or if he genuinely believes that the information security community has fewer than 1,000 competent experts.
If the former, yeah, the term "cybersecurity expert" is unfortunate - but it's clear it's just PR speak for "information security professional". Cringely then attempts to define that first, largely meaningless term, and then polls his anonymous friends (who themselves probably do not fall within that definition) to come up with wild guessess.
If the latter, yes, we definitely have more than 1,000 security experts. There is something around 500 emitent, internationally recognized folks publishing books, research, and otherwise contributing to the "cutting edge" of the industry. Then there's another 500-1,000 top-tier, notable security VPs, CEOs, etc, working for Fortune 500 companies (they may not all be technically savvy, but they *are* the industry). Then, there is probably something close to 200,000 security professionals working for companies around the world - we have something like 50,000 registered CISSPs alone (which is a certification largely inaccessible to hobbyists, and pursued by a minority of infosec workers), something around 50,000 subscribers to BUGTRAQ and other security mailing lists, etc.
Does this mean that DHS would be able to hire 1,000 competent experts? Unlikely, as the government historically did a pretty poor job of competing with commercial corporations (in terms of compensation and work culture), and many agencies may lack the hiring rigor and expertise to make the right calls. Given the size of the networked infrastructure in the US, this number is high, but does not sound outlandish by itself, though (many large corporations have 20-100 security people on their payroll).
With no other specific information released to the public it is difficult to say if this report merely reiterates one of these problems, or discusses a new vector; but regardless of this, it is a well-understood property that users sadly need to live with for the foreseeable future.
Most of the information security consulting companies are relatively small shops (5-50 people is common) with a handful of customers each. There is also a number of security testing divisions attached to some of the largest all-around international consulting firms, but they are relied upon primarily for regulatory compliance needs (meaning: "let's get this over with as soon as possible"), and they usually combine lack of any identifiable infosec talent with outrageous pricing.
So, with small companies serving non-overlapping groups of customers, it is almost guaranteed that no Slashdotter (of whom only a small fraction deals with information security!) can offer a meaningful, first-hand comparison of the services of key players in the field - and even if this is incorrect, there is absolutely no guarantee that the person telling you about their experiences would in fact have a sufficiently advanced understanding of computer security to make the comparison meaningful.
Unless you have enough in-house expertise and set up some controlled experiments, it's very difficult to tell if a positive outcome of a security audit means you are in the clear, or simply that the auditors are incompetent. To make things worse, even observing that auditor A identified n bugs in the setting in which auditor B identified n+m does not really tell you much, unless you truly understand their impact in the context of your services, or the reporting granularity and thresholds used.
What else? Many of the small companies may rely on PR alone, and some might be outright dishonest, for example releasing inflated security research, or simply astroturfing on Slashdot or elsewhere. And some might be run by people with actual credibility in the industry, but running subpar businesses because of poor project or team management skills. Just because they present at Black Hat, post to BUGTRAQ, or have a book published, does not mean a lot (but is a positive factor, of course).
So there's no easy solution. What you need to do is not to rely on Slashdot to give you answers, and instead, collect all the names you can easily find on the web (and in responses to this thread), then spend several days going through all the freely available primers on web application security... and come up with a decent RFQ that inquiries all the companies about their credentials, methodologies, the tools they use, sample reports they provide, and so forth. Ask technical questions, and expect them to be answered by technical people. You then need to set your bullsh*t detector to overdrive, and be wary of vague, dismissive, or nonsensical responses that look as if written by a marketing drone.
Based on this information, you then need to make the call which one would suit your business best. Good luck. It's not easy.
"Indecision is the basis of flexibility" -- button at a Science Fiction convention.