Yes, it is normal, especially at the salary that we pay (currently $275K). To take but one domain: at that salary, you are (roughly) at the level of a school superintendent for a school district with thousands of students (like, large town/small city). If you have ever been associated with hiring someone in that kind of position, you will see a VERY time intensive process, with many hours of interviews with many stakeholders. So yes, we have an intensive process -- but we take hiring very seriously, and we believe that it shows in our extraordinary team.
Oh how I wish you guys would hire new grads. I got into kernel dev partly due to zfs and partly due to Redox OS. Well atleast I have worked with Redox now, hopefully someday with you guys too :)
Its simple. Go to Google then. FAANG gang is decades older than Oxide and 3-4 order of magnitude higher revenue. They can hire hundreds of people a year who never do an hour of useful work in a year and still be fine. A startup can't do that.
I don't understand how these arguments about normality are supposed to be persuasive when so many of us have had the experience of being required to travel to do interview cycles. Nine hours does not actually sound like all that much to me.
Travel time is not equivalent to interview time. I would rather travel 2 days for a 3 hour interview than have a 9 hour interview in my current city. Interviews are exhausting if they are thorough.
> Nine hours does not actually sound like all that much to me.
I think because of who you are you haven’t actually had to experience a regular tech interview in 10+ years.
That's directionally true but note that I'm not talking about current interview practice; I'm comparing claims about current norms to what they were before.
(Though: I actually did interview somewhere recently.)
Well, bullshit actually. When I went to Sun in 1996, I was flown out for a full day of interviews (9 hours!) plus both lunch and dinner. (And this was true for more or less every company I interviewed with.)
I'm not sure why people are indexing so hard on the interviews, as we matriculate very few candidates into conversations; if you want to complain about something, complain about the written nature of our hiring process.[0]
btw congrats on the company. i was skeptical at the time but it was never because you wouldn't attract awesome talent and make great hires. your process looks great to me.
Thanks! And don't worry -- there were PLENTY of reasons for skepticism in 2019! We got lucky many times over, hitting on stuff that turned out to be really important in ways that we couldn't know at the time.
The hiring process is a very concrete example of that. What still blows my mind (at least a little) is that our process came directly out of one of the worst moments in my career: making an absolutely terrible hire about two years before we started Oxide. For a while after that, I wished I could go back and warn my previous self to not make the hire that would be so destructive to the organization -- but in the years at Oxide I have come to realize that that terrible hire was in fact essential for the future Oxide: it was the impetus for putting together a writing-intensive process, and there is no doubt in my mind that I had I not made this atrocious hire, I would have continued to hire the (broken) way that I had been hiring. (I described this a bit in my conversation with Gergely Orosz on hiring.[0])
Anyway: don't get me near the time machine -- I'll only screw things up.
>is that our process came directly out of one of the worst moments in my career: making an absolutely terrible hire ... that would be so destructive to the organization
What role was this hire filling? Are they publicly well known enough to name drop?
I think people are just generally grousing about Interview length, not about your company specifically. Don't worry about it. It's not aimed directly at you.
Well, in this case it was actually literally paying our taxes. (In the photo, Steve is about to initiate a wire to the IRS.) At our size, it's not a check -- and certainly not a credit card.
As to why we're paying taxes: while tax minimization is certainly a thing (and we have taken advantage of things like R&D tax credits), we're paying taxes because we're profitable.
I guess the question is why not invest those profits to further expand when there are already a backlog of orders to fill.
But on the other hand Huge Congrats being a profitable company! Hardware business is insanely brutal, and having the backlog is great to hear. AMD being investor I hope you guys get Zen 6c and Zen 7c earlier in the future.
A few ideas on why reducing profit by investing into the company might not be their best choice without the new investment round:
Hiring people is hard, takes time, and timing of big sales closing causing income may not line up well enough with the hiring process to offset that income fully.
Their machines are sold for high prices. It’s not small and not cheap. If a buyer pays in a lump sum that can be a big income and Oxide may not be able to time that ideally for tax purposes.
Buying electronics components is funny. Want to buy 1000 units? Price X. Want to buy 100,000 units, price is X/25. But till you can place an order for every component at the 100,000 quantity, you can’t get those price breaks so your cost to manufacture machines is much higher. $400M may let them buy at those bigger quantities and get those price breaks and it may not have been worthwhile to buy inventory at the lower quantity pricing if the investment round was looking to close real soon now. Cash flow is king, and hard, for hardware companies.
You're snidely dismissing the obstacles as being restricted to _today_, projecting into an arbitrary future to the point of being science fiction, and then complaining about "smug hand-waving"?
I don't understand why every person who takes the counterposition against you, Bryan, is always just talking about some hypothetical future like they're some Vedic astrologer, lecturing you about how this "real" future they see totally renders your argument pointless.
Your article is supposedly about "listening to the subject matter experts", but you failed to do that, and you immediately fell back to "all of your possible futures that sound really bad are science fiction with no possibility, source: me. My future prediction is grounded truth and guaranteed, source: me." Not that there's any meaningful future extrapolation about hacking anymore, but whatever.
If that's the only argument you're willing to defend, at least be up front about that.
The students in the computer lab had no guarantee that they wouldn't, say, download a copy of Napster infected with the CIH virus later. The fact that they were not under imminent threat from some kind of Hollywood-style network worm did not mean they were immune from more realistic attack vectors.
Likewise, one can quite reasonably say there is no credible existential, Hollywood-style threat from AI in the foreseeable future while recognizing far lower-stakes, yet important risks that need to be addressed.
the earth has been 1 decision away from explosion for nearly a hundred years now. do you anticipate a change to this agenda very soon? personally I imagine the same trajectory continuing.
certainly I behave like the article's author and yourself all the time, this isn't meant to be some sort of moral point. watching all these smart people talk so confidently regarding things nobody knows about, it reminds of the confidence ai shows in hallucination. would you say that the singular choice of vasili arkhipov did not prevent annihilation?
> would you say that the singular choice of vasili arkhipov did not prevent annihilation?
He prevented a catastrophe, but not annihilation. We were not at risk of that in 1962 and even if he had decided to go with the others, more decisions would have been needed (not just his) to fully escalate to full scale nuclear war. I will reiterate: We have never been one decision away from full scale nuclear war.
But in case you don't understand why, it's because no one person can actually launch all the missiles. And considering the two major arsenals (US and USSR), there has never been a time when two people could make the same decision (launch) and actually launch all the missiles. The orders still have to go out and acted on, many decisions have to be made in order to have full scale nuclear war and come close to annihilation.
I like the way you frame your opinion as a truth with an obligation to be understood. still I disagree that a decision can only be said to affect its direct successors.
If you think AI will be powerful enough to change the world in huge ways, then it does seem reasonable to assume some level of risk of destroying the world too.
Stating that the probability is zero when we simply don't know what the probabilities are does seem like an extraordinary claim.
Imagine how reassuring it would be to people if we had evidence that there's no existential risk?
> If you think AI will be powerful enough to change the world in huge ways, then it does seem reasonable to assume some level of risk of destroying the world too.
There's an opportunity cost in the doomsday prophesying, though.
The media coverage of "these extremely capable robots might kill us (according to the guys who sell the robots)" comes at the cost of coverage of the real, present issues surrounding the tech oligarchs that we're already facing.
Sure, there are some failure modes that lead to some really bad stuff, but human extinction seems exceedingly unlikely to be one of them, and "10%" is pulled straight out of Dario's derriere.
Far more likely are economic disruptions that impact the tenuous balance between labor and capital and lead to unpredictable societal upheaval.
No it does not seem like an extraordinary claim, unless this is your first time hearing such claims. For the rest of us, we've heard this every decade and it turns out to be entirely untrue.
Nuclear, Overpopulation, Peak Oil, Y2K Bug, Global Warming.
No, you explained why the selection bias doesn't matter, which I agree. The self-report data still is completely worthless in the first place though.
People are really, really good at lying to themselves, let alone to an online poll. They'll tell you that they prefer imperfect or even bad writing as long as it's not AI slop, just like how they'll tell you they like healthier food, they prioritize personality instead of look for potential dates, how they use LLM "only as a spellchecker", and how they use tiktok for educational videos. As long as there is no stake, people will just say what make they feel better.
Self-reporting data for human behavior is just noise.
I mean, fine, it's anecdotal -- but the numbers are also overwhelming (and don't seem to disagree with the tenor here).
And look: you're obviously free to ignore me because you feel that the survey data is "completely worthless" and just slop your way to success -- all I'm doing is trying to explain why you shouldn't expect me (and people like me) to read what you create.
Interesting! Amusingly, if I feed that detector this blog post, it identifies it as confidently robot (97 out of 100 test passages). And running through my last five blog entries, they are all over the map, with three deemed at least "likely robot." Looking further back in time (and taking a somewhat random example), a blog entry from 2008, "Concurrency's Shysters"[0], is also deemed as similarly confidently robot (also 97 out of 100); do you expect this high a false positive rate?
My experience is using Pangram quite often with lots of writing of all flavors (including a bunch of known origin).
As for my own writing, I didn't do this experiment, but one of my co-workers did -- and over 176 posts spanning 22 years, all 176 (well, 177 now with my latest) are 100% human. This is not hugely surprising in that (in addition to me having actually written them!) my voice is very... distinctive. What would be more entertaining would be to try to get an LLM to write like me and fool Pangram that way. I still think that this would be difficult based on the experiences that I've heard, but it wouldn't surprise me if you could pull it off (and I would assuredly find the result entertaining!).
In the dimensions that we use Pangram in the most actionable sense (namely, to audit our own public writing), I am unconcerned about false positives, and leave it to Oxide authors to rework/recast as needed. (Though it sounds like Freddie didn't even need to do that -- he just needed to provide a longer sample.)
Now that you’ve seen it can be brittle (e.g. if a small sample is provided, per this single case), would it be sensible to add a disclaimer to the post? It’s a great ad for the tool (& I’d love for a perfect tool to exist!), so it’ll sell subscriptions & we wanna make sure that some teacher out there doesn’t falsely accuse a kid, or engineer doesn’t think worse of their colleague unfairly, etc.
False negatives are mentioned, but the false positive is what could hurt people.
Well, give it a shot -- you'll likely find that that technique doesn't work nearly as well (at least with Pangram 4) as you think it might. When we had Max on the podcast[0], Adam explicitly asked him about exactly this (after all, you can give an LLM access to Pangram and let it iterate!), and Max reported that someone had attempted to do this -- and ended up burning through $700 in tokens and had a "sad Claude." Another interesting bit: according to Max, newer models are diverging more from human writing not less. I think that that was more anecdotal than quantified, but an interesting comment nonetheless.
99% of college essays and pretty much everything “product” in corporate America is now LLM generated with some marginal oversight. It passes muster for the most part.
This is (obviously?) false, but considering how well-capitalized we are at the moment, you do have me wondering what a quid pro quo would be for; perhaps in this fictional universe Pangram has lucked into some of the PCIe clock buffers that we've been scrambling to secure enough of?
By "quid pro quo" I wasn't suggesting that Pangram's PR people's podcast placement was pay-for-play, just that they traded access for your positioning of their tech and their exec in your content marketing efforts.
That's pretty normal, but the point is that a blog post which is 35% Pangram promotion may not actually be less annoying than the use of AI to help write blog posts.
Yeah, fair -- and definitely not: I am earnestly just a fan of what they built (and I also think it's really important as a way of getting a check against rampant LLM use).