dpifke

Location: Anywhere in U.S.

Willing to relocate: Temporarily

Technologies:

I’m looking for 3- to 12-month contracting opportunities (corp-to-corp), anywhere in the U.S., with organizations looking to move some or all of their infrastructure on-prem.

I can...

Design reliable, secure, scalable, and cost-effective compute, storage, and networking that’s tailored to the needs of your business, using open source software, open weight LLMs, and off-the-shelf hardware.

Find suitable colocation host(s) and negotiate agreements for space, power, bandwidth, and insurance; or hire and supervise a licensed general contractor for building out a server room in your office or other facility.

Get the best deals on equipment, cabling, transportation, and everything else.

Hire and supervise temporary labor for the buildout (“rack and stack”).

Bootstrap networking, storage, cluster management, and Gitops systems. I have lots of experience running Kubernetes on bare metal, arranging BGP peering at public interconnection points, and connecting private clouds to public ones.

Perform initial validation and load testing, and help you migrate services from public clouds or SaaS providers.

Provide training and consultation on SRE best practices, observability, oncall rota, etc.

Resume/CV:

I have 30 years of experience in ISP, web, and related businesses, including as a Technical Program Manager for Google SRE. I started an ISP in 1996, back when all that was needed to do so was a bank of modems and a Linux box or two. I led data center buildouts for several dot coms—including a major social networking site, Bebo—in the late 90s and early 2000s, before the “cloud” existed, and I have tons of experience in the intervening years writing software and operating production services on all of the major hyperscalers. I had a stint managing wifi buildouts in hotels all over the country, where I learned the ins and outs of finding and working with skilled and reliable tradespeople. As the founder of two startups, I know how to be scrappy and move fast in a high-growth environment; I’ve also been an employee and consultant embedded in more mature organizations, always with a reputation for getting things done.

Email: dave at pifke dot org

I've interviewed hundreds (maybe thousands at this point?) of candidates in the course of my career, and as much as I think giving feedback at the end of the call is a great idea, I personally would struggle with doing it in a way that was actually useful to the interviewee. Because unless it's a hard "no" (e.g. someone is obviously lying about their experience), I almost always need to ruminate on the interview afterwards before I can write up useful feedback. Sometimes it's the act of writing itself that helps me form a coherent recommendation.

This is also why I hate interview processes where my co-interviewers want to debrief immediately, or when I have a recruiter breathing down my neck as soon as the call ends.

In this case, "null MX record" means MX exists, but does not specify a valid server:

   $ host -t mx example.com
   example.com mail is handled by 0 .

Senders should not fall back on the A record in this case.

I use uBlock Origin for this, something like:

  news.ycombinator.com##:matches-path(/^/item\?id=/) tr a.hnuser:has-text(/^dpifke$/):upward(tr)

This mostly works, but only kills the user's comments and not replies, so it sometimes can be confusing.

The very first subheading is entitled "What to Submit." I quoted it in my initial reply as rationale for why the people flagging this submission as off-topic were justified.

This seems similar to the "Is Github Down?" submission problem, where the submitter simply links to github.com.

That's a poor submission, because by the time most people click on it, Github will no longer be down.

There might be an interesting discussion to be had about outages at Github, but the better submission would be an article or blog post about the outage, not just a link to the site and a three-word title.

If someone wants to write an article or blog post about this news broadcast, which links to "hard facts and analysis not available through popular channels," that seems like it might be a worthwhile submission. But just a link to the broadcast by itself is not leading to interesting or on-topic conversation—the top comment right now is an ad hominem attack against Larry Ellison, without any supporting facts or analysis that he had anything to do with this story at all.

Lots of hackers find porn very interesting. In fact, my first "real job" as a hacker was for a company with ties to the 1-900 industry that had decided to expand out onto the internet (not just to sell porn). Stories about porn would be interesting, submissions of nothing but pornography itself ("because it's censored!") are not.

I would be more sympathetic to the argument that this is relevant if the submission was an article about media censorship, or CBS's audience or leadership, and how said censorship, audience, or leadership relates to technology or emerging trends in media.

But this is literally just a controversial TV news broadcast, that people of one political persuasion say was "censored" and people of another political persuasion say was held off the air "temporarily" until it met network fact-checking standards. That sort of political bickering is most uninteresting, and is most definitely not why I've been reading HN for the past few decades.

Strong disagree.

Gay porn is censored in a lot of Muslim countries, that doesn't make it on-topic for HN.