int0x2e

A rumor I heard somewhere - MS supposedly built x86 emulation for ARM a few years back, and created a demo on top of some vendor's ARM servers for transparent Azure x86 on ARM (in addition to regular ARM VM SKUs of course). But rather than move forward with this as a public offering on Azure, they used it to put the squeeze on Intel/AMD, and they got rock-bottom price on hardware in return for holding-off on ARM adoption.

Sounds like someone may have gotten a massive bonus that year, but may have significantly delayed any ARM efforts on Azure...

"we'll never run out of lithium" is the new "640k ought to be enough for anybody".

As we make lithium battery capacity more affordable, it gets applied more broadly, and the rate of use goes up. There is a very finite amount of any resource here on earth, and while we can get things more and more efficient, there will be a moment where we can no longer extract lithium (or other materials) required for battery production. Treating these as an endless resource is the same error we made with fossil fuels, just hopefully without as much climate change as a direct result...

In my experience, this only works if you have a small team of great engineers. Once you have 50+ people of various skill levels making contributions to the codebase, you'll start seeing your git history littered with variants of "fix", "fix 2", "fix broken CI test" and on and on...

For me, the PR is the context I can use when trying to find an issue. Sifting through many hundreds of commits per day is painful. Sifting through tens of PRs per day is not great, but is much more manageable...

I often feel the same way. I've seen cases where trying to add functionality to an existing but "very clean and correct" codebase ended up being extremely difficult and the result felt far too complex in my mind.

I've come to think the real sin is coupling. If you try to keep your codebase clean at all times but don't refractor often enough, you end up with coupling where it shouldn't be, which can sometimes be a lot worse than some of the so called "deadly sins" such as code duplication.

One thing I think can really help is to have two different modes within the same team. You move fast and cut all corners and rapidly iterate when you have an idea by building a prototype, with the only goal being maximizing the speed at which you're learning if/what works and what does. Then you basically throw away all that prototype code and build the feature properly with careful attention to code quality and complexity, testing (++) - since now you actually know what you're building and that it's likely a good idea.

Rinse and repeat, so that you only have prototype code where you're actively learning, to be immediately replaced by properly engineered solutions or wiped away (with the same haste you used to create it initially - no "rarely used, badly engineered" features should stay in your codebase due to priority / resources / attention constraints).

The biggest challenge I've had with this approach has been building and fostering the trust and culture needed to have everyone truly go "all in" for both aspects/modes of working. This breaks down quickly if you have managers or PMs drive engineering to keep the prototype code in for longer, or if engineering insists on not cutting corners during the exploration/prototyping phase...

You hit the nail on the head - at many places, the hiring process seems to be great at blocking people who are great at solving real world problems that aren't as good at whiteboard leet-coding.

If you truly enjoy this part of the job, you can go for a freelance career - build a network and reputation as the person you call to solve the hard stuff, and do mostly that, no leetcode interviews required...

In a document store, what does it even mean to create a container without a partition key? The document store has to partition the data somehow, and doing so implicitly sounds dangerous to me since all you're doing is creating a hotspot on one of the partitions...

Cosmos != CosmosDB.

The two have nothing in common (and trust me, it sure is fun having to constantly make sure which of the two someone is actually referring to every time...).

I absolutely want to work in a diverse team. I want to be challenged by different perspectives and ideas. I want to have the best team members out there, regardless of their backgrounds.

I however, think many D&I ideas fail to work when written into policies, and I think many times, higher-ups seem to write policy out of the best of intentions, yet fail to see how they can easily lead to abuse and poor results.

Two examples from a corporate job a friend shared with me -

1. A friend was interviewing senior software engineers for an open position, then got a candidate who was grossly irrelevant (I won't bore you with the details, but she has ~1 year of experience at best, and this position was targeting a minimum of 5 years, 8+ years preferred). Turns out, that specific hiring manager had recently lost a (male) top candidate because they didn't have a female candidate in that candidate's pipeline - the manager learned his lesson and now always dumps 1-2 female candidates that have no chance of making it into any senior position, just to meet the "diverse slate" requirement of their D&I policy.

2. Another example, this time not related to diversity - during the recent economic downturn, a company decided that on top of an aggressive hiring freeze, for any employee that is fired or quits, their position goes away with them, then gets re-assigned to where the higher-ups think the need is greatest. At first glance, that sounds like a good idea - kind of a CI/CD for re-orgs with minimal "slack". But of course, if you're a smart enough manager, that means your biggest hope is to remain static through this period, which means you'll never fire (or even challenge) anyone. You'll even promote mediocre folks to keep them on, then fire them afterwards.

Naturally, the most talented will still get plenty of offers and move / leave at some point, and then the slightly less talented folks get hit with all of their work and eventually move on as well, and over time the average talent level of that team will slide down considerably, and as we all know - hiring a top performer into a mediocre team is a challenge.

Nice policy in theory, but in practice - it will cause the company to lose tons of great talent, in a way that will take years to recover from.

I actually think pixel density is one of the easier battle to win. Getting a cheap, lightweight, strain-free device to users is the real requirement.

For now, I can't handle more than ~30 mins with any VR/MR/AR device, but if I can actually wear one for several hours and be productive using one - I will be convinced we've turned the corner on this tech.

If you treat the homeless as lepers and stick them at the edge of town - you will get exactly what you said. If however, you create one home for the homeless on every street/block/floor - you'll find out that they are people, just like you and me, and with a little bit of support many of them will even thrive. The US is just so big on NIMBY that there is zero chance of that happening :-(

You might be interested in listening/reading a recent Planet Money episode (https://www.npr.org/2022/09/22/1124477182/federal-judges-eco...), which explains how for ~20 years or so, many US federal court justices went to an exclusive, all expense paid economics bootcamp, taught by some of the most well known economics thinkers of the time. That sure sounds great in theory - why wouldn't you want to have judges to have a deeper understanding of economics when they can have impact on financial matters?

Well, as it turns out, some of the key ideas taught focused on the rational economic theory, and extrapolated from there to crime and punishment - e.g.: a criminal makes a cold economic calculation before commit each crime, taking into account the expected gain from this crime, the likelihood of getting caught, the expected punishment if caught, etc. - and therefor, if we make the "cost" higher, by opting for harsher prison sentences - crime will become less attractive and therefore happen less.

When the data is analyzed, justices that went through this program appear to hand out harsher sentences to offenders, and often rule in favor of corporations and against government agencies when compared to the control group. The effect is most evident for judges that used to learn more to the liberal side, which is surprising to me, but it definitely is interesting.

A former boss taught me of the "law of conservation of bullshit" :-)

I feel the same way, but to play devil's advocate - the vast majority of startups do fail, and leave you with no meaningful legacy.

You could grow as a person, you could be very passionate and driven, etc. - but you could also end up quite depressed when things die down (speaking from experience sadly).

I still think startups are awesome, but if you view work as a transaction, having a good corporate job can be a better transaction for many folks, especially if work is only a part of your life, not the defining feature. Time with your children when they are young is often the easiest to explain, but same goes for time with parents/grandparents, time to work on passion projects, etc.

Google is known for launching and then soon killing its projects, and game platforms are built on long term trust and momentum. That's why Stadia never got the momentum it needed.

Feature flags or another way to release bleeding edge features to only some of your users can sometimes be the best way to iterate over new features quickly when your most user base is risk-averse

Adding my own personal story as an anecdote - I never completed my bachelor's (double major in physics and CS). I did okay on my physics classes, great on my CS classes (often in the top 2%), yet failed badly on a couple of advanced calculus classes. It got to the point where I was done with most of the requirement and elective classes on the CS and physics side, was already taking graduate courses in CS and lecturing in a course I helped create in CS - yet failed out of my bachelor's.

In the end, the industry offers for me to go all into tech got too good, and I managed to advance to middle management with no need for that degree, so I never went back to finish it. I still think academics are important and I am thankful for the stuff I learned there, but I think the fact calculus and "hard maths" are such a strong focal point early on serves no purpose other than serving as an arbitrary filter for most people.

I suspect it will never happen, but I would love to work at a company that has front-line managers and recognized "thought leaders", and nothing else at all, most of all - no middle managers, little to no concept of "promo work", etc.

Once you remove the incentive for managers to grow their org (because there's a point where managing more people directly becomes infeasible, or at the very least - painful), you'll have little to no politics, no more constant re-orgs, etc.

Wishful thinking, I know...

Lots of great insight. Thank you for sharing!

Your comment about being too early is on point, though I'm not entirely sure it's a bad thing. If you're a big company with scale and cash to support it, exploring new frontiers is probably worth it - you either stumble upon the next big thing, or you at least get to stake your flag through some patents, and in the process - make it so if this things does grow - you already have some kernel you can build upon to start that up without having to pay up for an acquisition...

It's not perfect, but from a risk / reward perspective - I can see how this strategy could serve a large tech company well over decades...

If I was designing this, I would only have a small set of "SKUs" with minimal customization, and then go all out on economies of scale.

If 10% of houses switched to a single "system" with a small set of SKUs, such that everything is optmized for manufacturing, shipping and assembly - you could reduce the cost of construction significantly. This is basically an extension of the IKEA model for the house itself instead of just the furniture.

I would argue moving would require moving the VC industry as well, but if anyone can do it, YC would be the one to get it done. Having said that, it's going to be a real challenge - SV's prominence is quite massive and long-held. Founders would need this new location to offer similar benefits in terms of talent, investor access, etc. to make it worth their while.

Just curious - why do assume a correlation (one way or the other) between these two metrics?

Apple is super popular in the US, but in the rest of the world, and especially in non western countries, Google (Android) and Samsung have a much larger network...

The issue is that domains are cheap and plentiful, and a spammer can easily claim to be sending mailflow on behalf of let's say, 1000 users, each sending 5-10 emails a day, and thus, a single IP "forwarding" (sending) mailflow "on behalf" of 10k domains, with 1k users each, with each user sending just 10 emails a day - would result in a pretty nice batch of spam making it through.

As things currently stand, the only "expensive" resource that is hard to acquire or fake somehow is source IPs with good reputation. Fair? I'm not saying it is. It's just effective.

While I agree with you generally, one could make the argument that if your aim is to drastically disrupt some domain, you may wish to adopt a riskier but faster path even if you're working in a safety-critical field, because the "slow and safe" route means more lives lost / negatively impacted while you slowly perfect your solution.

If you'd humor my extreme utilitarian view for a minute, I would argue both autonomous self driving and some medical endeavors could save many more lives if we, as a society, said "you get a budget of 10k. 10k lives you can severely negatively impact to deliver impact greater than that number" - basically giving you an investment/debt you repay to society via your future impact. Currently, traffic fatalities are at ~38k/year in the US (with over 2M/year injured), and the numbers for leading medical causes are staggering (heart disease ~700k/year, cancer ~600k/year, etc.). I would argue our current processes for breakthroughs in areas where health or safety are involved simply lean too far towards the "safe" end of the spectrum.

One anecdote I can share is a friend who worked for over a decade on a system where patients could buy a medical test at a pharmacy, take a urine sample at home, and get lab quality results using their phone's camera. The tech was ready in their first year of running. They built a suite of validations and tested things across hundreds of phone models, they really did their part well because they truly care. Getting things FDA approved and in patients' hands took much longer, because the processes are extremally slow and designed to reduce risk by almost all means necessary. While that's a good idea in theory, it didn't stop a scam like Theranos (because when you're intentionally dishonest rules don't always help), and it did make it so my friend's company took a lot longer before being able to get fast, accurate, test results for many different metrics to rural and poor communities where lab testing can be an issue...

I suspect you haven't met too many people in these programs. I had a student in one of my university programs who was handicapped and could barely see 3ft in front of him. Guy had to ride in a motorized chair, get guided around campus, and use a super-zoom lens and a laptop screen that made everything 10x-50x its size to be able to read the whiteboard/presentations. I'm sure none of that stuff was cheap.

But that guy was awesome, fun, and super clever. His lowest score was an A-, and he had a great character. I'm sure he had to work ten times harder than I ever did just to get into that position, but I never heard him complain. Personally, I think keeping a mind as great as his in some sort of hedonic trance instead of letting him learn and contribute would have been a great loss, and possibly quite cruel.

I totally agree with you from a usability standpoint, and while it is possible to make this work well, for larger-scale services it is rarely without costs, and for the most part - people don't seem to care as much as you'd think.

While it can be frustrating, I doubt much revenue (relatively speaking) was lost due to this issue, which means for most apps and services, it will likely not get done.

(I'm not disputing the merit of your arguments, just explaining why this will rarely get done well in real life...)

Bryan Cantrill's talks are some the best I've ever seen. I've always tried sharing them around with colleagues (with limited success, but still worth it in my view...)

I couldn't agree with you more. A FAANG-tier company with which I'm closely familiar pretty much did the worst thing every time during the past 2 years - they tried to tighten the belt when COVID first hit (missing a ton of great talent when the initial uncertainly killed off some of the smaller companies), then they tried growing and chasing talent really hard, right when the bubble was at its peak, and now that the economic outlook is negative, they've halted / significantly slowed down hiring. My personal opinion is that Warren Buffett's old quote to be "fearful when others are greedy, and greedy when others are fearful" rings really true.

For Starsky's engineering / demo purposes - they could limit themselves to a few safe routes with good connectivity. Just use a couple of cell modems from different networks, maybe add a satellite backup to be extra safe and you should be safe enough if the truck can do basic collision avoidance and lane-keeping.

For the future, the good news is that the world is already headed towards a state where roads come with high-speed connectivity infrastructure for vehicle-to-vehicle, vehicle-to-infrastructure and vehicle-to-x communication. Building reliable, high-speed connectivity to the road makes a lot of sense if you want autonomy to really take off.