HNHacker News
TopNewBestAskShowJobs

nhoughto

451 karma · joined October 9, 2014

submissionscomments
nhoughto··on Newegg has a bit of a scandal on its hands
Is Newegg bad now? Historically heard good things, what happened? Private equity?
nhoughto··on Nearly a third of TV ads play to empty rooms
It is interesting to me that TV off is the norm (for me and people I know) except! If there is a live sporting event on that someone cares about, then it would be rude not to have it on and is expected. Weird. Makes no sense at all, humans are weird.
nhoughto··on Resources to encourage children to learn to code
I look at the self directed learning resources kids have today and am jealous. How many of us have brute forced computers and the abstractions within by hammering MSDOS, all of its weird nooks and crannies with no one to explain anything.

I don’t know what I’d do if I could just YouTube any topic and get a meaningful answer 30 years ago, it could be incredible? Would have made things faster but maybe that’s not the point? I’ll find out I guess with my son, it will be interesting.

nhoughto··on I would like a job writing Haskell
I wonder if it’s a signal of how much their are willing to Invest in people. I’ve been interviewing recently and have seen many companies are explicitly “we don’t care about language exp we will teach you, we want good ppl”.
nhoughto··on Building a modern home in the woods
Interesting write up, a not comparable example as my house build was nothing like this scale or cost (I imagine), but during building our house it was interesting to observe where problems came from and how they could be fixed.

The architects -> engineering -> builder handoff is obviously the most lossy, tons of context/intent/purpose is lost in these steps, and normal way incentives align doesn't help with each group being sympathetic to the greater mission, the more they care or clarify etc the more time things take / cost incurred. It is natural to assume everyone is thinking the same as you and will be motivated because you are paying them, but they rarely are, they care less about the result than you do so staying on top of the details you care about throughout the process is a must.

What worked great for us was to have the builder part of the architecture design process and inject their learnings of how to actually build the thing along the way, rather than trying to bolt it on at the end. This isn't normally the way things work as normally architectural drawings are a primary input to a tender process to find a builder.. so is a bit cart before horse but we already had a builder so could avoid that.

nhoughto··on Wormhole confirms all $320M in funds have been restored
Restored is an interesting choice of word, Jump replaced the ETH at their cost is my read? It’s not like they wound back the transaction or got the ETH back.

They would want to be confident there are no more bugs, only a few days ago this happened, did they do a full audit of things before tipping it in? Imagine if they lost another 320m !

nhoughto··on How to stop running out of ephemeral ports and love long-lived connections
28000 ephemeral ports is enough for the c10k problem, but 100k+ seems a stretch. At what point is increasing the number of ports from 28k to a higher number the right answer? Reuse as described here sounds like a useful optimization but at some point (or due to pathological workload) even that will be exhausted I’d think? What to do then
nhoughto··on The Road to OCIv2 Images: What's Wrong with Tar? (2019)
Squashfs instead! There was another post about this recently, replace tar with squashfs solves a bunch of problems. Is arguably harder to create and had a smaller ecosystem than tar, but capability wise it would be great.
nhoughto··on Show HN: Add a “1” to your Hacker News URL to get fancy social previews
Handy, I’ll probably forget the 1 suffix tho. Ycombinator-meta.com or smth might be more memorable, that’s my problem tho I guess =)
nhoughto··on Show HN: Add a “1” to your Hacker News URL to get fancy social previews
You could probably just do both? Once a redirect has fired, off you go..?
nhoughto··on Speeding up LXC container pull by up to 3x
good fit for containerd then =)
nhoughto··on Speeding up LXC container pull by up to 3x
Have thought about writing the same thing, would love to see it OSS.

Real value for me would be to integrate it with containerd or similar to speed up image layer pull times, I imagine that isn’t shelling out to a separate binary tho.

What language is it implemented in?

nhoughto··on Speeding up LXC container pull by up to 3x
Came here to say exactly that, squashfs with overlay mount for writes beats this easily.

Hard to change the entire container ecosystem away from tarballs tho so this is still useful. Is it available anywhere? I didn’t see a link?

nhoughto··on Ask HN: How do you explain the sloppiness of modern software?
Thats why we got AutoRecovery as a feature! /s
nhoughto··on The rising human cost of sports betting
Nah they generally make a loss on just the pool seeding bets (made by the nominated lucky few) and make it up from the rest. Generally the seeding bets end up as a small percentage of the overall pool.
nhoughto··on 1k-cycle lithium-sulfur battery could increase electric vehicle ranges
I don't have an EV but is in my future, and we do a long trip (500km) regularly, not sure how practical this is but i've always thought, is it worth buying a car for 100% of your needs? or get something for 90% and solve the 10% differently by renting?

Buy and use an EV for all the easy daily stuff, then rent something more suitable for the yearly/less frequent long trip. Other than range we have a similar problem with space, for big trips we end up with tons of stuff and more space is useful but day to day less so. So the 90% EV could be smaller as well and still solve the problem.

I do see the attraction of having one thing that solves 100% of the uses though.

nhoughto··on The rising human cost of sports betting
Yep, few friends gone through the same thing. Seems like a coder-finding-betting rite of passage, 'oh right i can automate this process and earn whilst I sleep!'

Nope, bookies only want to keep profitable customers, or provide some other upside to them. In the context of bookies they have limited other ways to make money but sometimes their are weird edges that pop-up, like legislated restrictions.

In Australia those running the licensed tote (totaliser) market pools on horse racing can't make the market themselves, they can't legally seed a pool, and since a tote win is a portion of the pool total customers are less inclined to bet on a low value pool, so the markets allow a few high value bettors to seed the pool for them (with generous rakebacks) which they make a net loss on, but has overall positive effect on their business.

nhoughto··on Pay attention to WebAssembly
yep was my thoughts too, WASM solves some problems well. But it isn't and can't replace general containers like Docker does, for some specific workloads sure but not a general solution. And yeah i don't quite think everyone talking about WASM understands its not just 'faster javascript' but a totally different world where you have to for example, BYO GC (or manage memory yourself if you like!).
nhoughto··on The new hire who showed up is not the same person we interviewed
reminds of me of this post, however true it is..

https://news.ycombinator.com/item?id=27454589

nhoughto··on People don't work as much as they say
Motivation isn’t fungible and I’ve been in a place where the given schedule is true of me, but that’s not a good place for me, that’s me unhappy and clock watching, a sign of bad times. I want to do more, solve more, continue . The end of day comes around quick.
nhoughto··on Can I have a smaller Prometheus
Interesting detail, didn’t realize go dead code analysis was this easily made ineffective. 52MB for ec2 SDK is a big chunk of code that must be mostly dead/unused.

Seems like eventually this should become a priority for library writers to support dead code analysis, otherwise going to get worse and worse..

nhoughto··on How the Soviets put a lander on Venus
Not surprising I suppose, likely the inspiration, but the pressures and temperature and general inhospitable-ness sound an awful lot like the alternate life supporting planet for Rocky in Project Hail Mary. Good book!
nhoughto··on Rancher Desktop 1.0
minor additional benefit is that it configures 'service account signing' in k3s OOTB, so you don't have to hack it in by changing config behind docker desktop's back if you really need it like we did.
nhoughto··on Rancher Desktop 1.0
Just in time for the end of jan Docker Desktop grace period (https://www.docker.com/blog/the-grace-period-for-the-docker-...).

I played with this when first announced, and then now and all the things I wanted when i first tried it out are now fixed (mainly transparent networking..) and the option for dockerd container engine makes migrating pretty simple.

nhoughto··on Google Drive flags file only containing “1” for copyright infringement
It’s a broken window. Is a perfect example of the craziness of the current state and approach that even the Everyman can understand and be surprised/question/have negative sentiment about.

The MBA thinking ignoring intangibles is exactly the problem, trading short term win for long term loss. Will it actually matter in the end? Probably not and the cycle will continue..

nhoughto··on Atlas – Terraform but for Database Migrations
Yep is a big long tail to think about solving for, decades of features / idiosyncrasies in a complex domain.
nhoughto··on Atlas – Terraform but for Database Migrations
Agree with general sentiment that creating a DSL/HCL like thing to essentially represent DDL commands seems less than ideal I wouldn’t want to be the author on the hook for never ending sql functions etc, you’d want a good escape hatch, same pain as hcl. But for a tool like this to add value (just like terraform) it’s got to understand the relationship graph, properties of resources etc.

I’m a big believer in manual simple sql migrations via flyway or similar, but if you got this working in theory and built the features it could do a better job on average than non expert humans, it could enforce patterns / naming / conventions / indexes etc. Could do with sql migrations too but would be more difficult as it wouldn’t have the full declared model.

nhoughto··on On Leaving Facebook
Was going to disagree but ended up coming around to the argument. I guess the difference is at a certain point preparation turns into a wasteful arms race, it’s not just a reasonable amount of effort put into your next gig but too much effort on things that don’t really represent real day to day / reality. Not sure where the line is, but there is definitely waste and it definitely favors those with spare time.

Does seem like there is a growing movement to explicitly not do leetcode style reviews at many places which is good.

nhoughto··on Efficient Pagination Using Deferred Joins
yep its more complicated than my simple statement makes it out =)

i guess the core of the statement was that the net cost to the db is more tied to offset than select, yes they interact and 'select * with offset' is more expensive than 'select foo with offset', but its really the offset that is the problem. If you did keyset style pagination and still did 'select *' you'd be fine and roughly the same cost as 'select foo'.

nhoughto··on Efficient Pagination Using Deferred Joins
Yeah it doesn’t, the mistake is thinking the “select *” is the expensive bit, for limit 15 it’s cheap, it’s the offset 150000 that is expensive deferred or not.
← PreviousPage 4 of 9Next →