I really doubt that. People have been hating on GitHub for years or even decades. GitLab at some point very publicly wanted to become "the world's most trusted place for open source software" but I think they gave up, or at least pivoted to AI.
GitHub has one massive advantage which is even people in HR know programmers use it, and they can just glance at a candidate from GitHub. For as long as this remains in place, GitHub will survive.
I would rather use GitLab, honestly. Forgejo, Codeberg, etc, have a CI/CD modelled after GitHub actions which I really don't like, but I digress.
Unfortunately that does not really work for people who live in countries governed by oppressive regimes, or people who are in any way different (immigrant, LGBT, etc), and in fact, even posting with the best of intentions will have people wanting you dead. Ask me how I know.
Actual inboxes with their separate username@domain and password? Probably about 20, out of which I only really use around 10 and the rest are dormant or I don't really bother checking.
It is not a bad post. I especially like that you point out at how AI is not really going away, in stark contrast to the comments I read daily on social media roughly saying "I can't wait for the AI bubble to burst".
Well, I am sorry, but the "AI bubble" is not really going away anytime soon, so you better get comfortable with it.
But indeed, it seems more important than ever now to build things and learn more, and not rely on AI excessively.
Because it blocks domain. The opposite of what a resolver should do.
It's a second time in ~2 months DNS4EU is reported on HN of not resolving domains.
Wait, what? Vivaldi is open source? Now I am confused and really not sure what was the reason I ignored it for so long. Was there something iffy with Linux desktop integration?
Quite surprised at Vivaldi. Considered that as Opera spiritual successor including any possible feature, will've been one of the first browsers adding AI.
Steve Jobs delivered presentations like absolutely nobody else, but for some reason, people continue to cram entire paragraphs into the slides. Boring. I wish people learned to make presentations. I have known how to properly give presentations since I was 12.
It seems this is little more than a funnel for us to buy your 130 USD book.
130 USD from a complete stranger is quite the ask. Especially because, as you mentioned, you don't even work at Google.
Your GitHub also does not have a lot of content beyond a few pointers which frankly does not inspire confidence in your project.
I understand you have possibly dedicated many hours to this, and I mean no disrespect, but I really have no reason to trust you. The 130 USD book could have been written by ChatGPT for all I know.
Fair feedback. I expect skepticism, especially given the price point and the noise in the interview prep market.
Regarding the 'ChatGPT' point: I’d argue the opposite. AI tools are great at generating generic definitions ('What is an inode?'), but they struggle with the specific operational sequencing required for NALS. For example, AI rarely suggests 'draining traffic to a fallback region' as a step before 'grepping logs' unless explicitly prompted. My focus is on that sequencing (The 'OODA Loop' for incidents), which comes from analyzing failure patterns in debriefs, not just scraping docs.
As for the GitHub repo: The goal was to open-source the core frameworks (The NALS Flowchart and the Linux Cheat Sheet) so they are useful without buying anything. If it feels too light, that’s on me—I’ll look at adding one of the full scenarios from the workbook to make it more standalone-valuable.
Thanks for the honest take.
GitHub has one massive advantage which is even people in HR know programmers use it, and they can just glance at a candidate from GitHub. For as long as this remains in place, GitHub will survive.
I would rather use GitLab, honestly. Forgejo, Codeberg, etc, have a CI/CD modelled after GitHub actions which I really don't like, but I digress.
reply