HN user

deodar

69 karma
Posts0
Comments73
View on HN
No posts found.

Drive failure rate versus age is a U-shaped curve. I wouldn't distrust a used drive with healthy performance and SMART parameters.

And you should use some form of redundancy/backups anyway. It's also a good idea to not use all disks from the same batch to avoid correlated failures.

All the cold data is on spinning rust, and most of the data at large scales is cold. Much of the warm data is still on HDDs too, fronted by a flash or RAM cache.

The storage densities are constantly increasing to keep costs low. SMR drives have significantly pushed up the achievable densities at the cost of write performance which matters little for cold data. The 1PB/node mark was breached a few years ago.

Few players have the budget to store these volumes of data on flash, especially since you're always erasure coding for redundancy which adds anywhere from 50-100% storage overhead, depending on other tradeoffs.

HDDs are a remarkable invention and not going anywhere soon. Heck, I think tapes aren't dead either.

A good pair of binoculars will be sufficient. You didn't need a ton of light gathering capability for casual planet viewing like this.

Telescopes are a bit of a rabbit hole. Many cheap mass market telescopes are also known as hobby killers. A 6" dobsonian (reflector) is a good starting point for deep space objects like nebulae and star clusters. For planetary viewing Schmidt-Cassegrain telescopes are great.

However, learning to use a telescope requires time and patience. Taking it to the field for an event like this for the first time may be frustrating as you will be spending most of the time figuring out how to collimate and align it.

I certainly don't mean to discourage you from getting one though.

A good pair of binoculars is much easier to use. They require no collimation out of the box and show an upright image that makes it much easier to navigate the sky, at the cost of reduced magnification and light gathering capability. You will be surprised how many celestial objects even 10x magnification reveals that are invisible to the naked eye.

Happy planet gazing!

My opinion is about the SF Bay area. Yes we have easy access to nature, year round good weather, and tech jobs are more plentiful compared to many other locations in the US (that's a lot of positives admittedly).

However the Bay area is also a cultural desert. There are no other major metros within a reasonable driving distance, and a paucity of museums and cultural activities in SF itself. Especially compared to living on the US east cost or even Los Angeles.

Distances are long, the municipal water quality is abysmal, everyone is hyper competitive and living in a bubble where being a millionaire is considered poor.

The road quality is not befitting a first world country, the cleanliness of streets and interchanges is below par, and public transport is a joke, enforcement of traffic laws is non existent, and taxation is ridiculously high relative to what the government invests in infrastructure. The California government is probably one of the most corrupt in the country.

I only live here because moving would be too disruptive to my kids. Once they go to college, I'm getting out of here.I just wanted to temper the jealousy sentiment. :)

I enjoy face to face discussions and whiteboarding sessions on tough technical topics. And I'm a fairly introverted person.

You get 2-3 competent engineers in the same room with no screens, just a whiteboard, and magic happens. No amount of zoom or virtual whiteboards will replicate that experience for me.

Also I don't learn much being remote. Many of my colleagues are 20 years or more younger than me and yet I learn something new every time I have deep discussions with them.

I'm lucky that my commute is 30 minutes each way. I choose to go to office at least thrice a week and a subset of my local colleagues do so too. If the commute was much longer, then I wouldn't go in as much.

Why Not Comments 2 years ago

There is no hope for the software industry to mature if we cannot agree on some basic coding practices. Like the judicious use of comments to improve maintainability.

I wish that were the case. I would have loved to have a more diverse pool of candidates apply. This was a mid-sized tech company (1500 employees) that is very well known in its vertical. We'd post job openings on Twitter, LinkedIn and the company website at least.

Yet the pool of US-based applicants was anything but diverse. Very few women and almost zero URM (Under-represented minorities). Also the candidate diversity was worse for more senior levels.

I agree.. I am sorry about the flippant comparison. It was probably unfair to scratch. My frustration with Scratch stems from watching kids (including my own) being taught it in school as the de-facto option and learning absolutely nothing about programming or anything useful in the process. Perhaps it is due to bad instruction.

For the amount of time and energy invested they would have been far better off learning a "regular" language like Python. Anyway just my 2c.

Victory has many fathers. Defeat is an orphan.

I guess that explains the number of people being touted as "The Godfather" of AI.

Definitely untrue. Like everywhere else it's a wide spectrum. You'd be surprised how many Indians come to the US from modest means. Parents sunk their savings into educating their kids and sending them abroad. It is not uncommon for these parents to now rely on their foreign resident children.

A lot of Indians (and I am sure it's true for other nationalities) owe a lot to the Tech industry for absolutely transforming their standard of living in the last 1-2 generations. Tech was also remarkably egalitarian since anyone with the inclination could learn it with a relatively modest investment in a computer or even unimpeded access to a computer at school.

I wonder if future generations will see the same lifestyle transformative effect. Or perhaps the new bar is access to an H100 GPU.

I recently switched to vscode after using vim for over 2 decades. Perhaps I am just getting older, however I now find that hitting 5j or 12k as in the article steals some cognitive bandwidth. I still use vim for looking at log files and quick edits e.g. to config files, and went mode-less for more free form editing like coding.

It may be slower and less efficient to hit the down key 5 times or alt+right a few times however it is also mindless and doesn't take any mental bandwidth. YMMV of course.

Not sure whether distracting is the right word however the internet is likely involved. Constant connectivity and engagement maximizing algorithms ensure they are exposed to the horrors of the world daily. It is difficult to imagine the existential angst this would be creating in their young minds.

I believe this is the "leaded gasoline" of our times. Something we will eventually correct for but at great cost to a generation of young people.

Leaving aside the technical merits of Golang's error handling, this article is a straw man. The Python example can be rewritten with error handling. And arguably in this example the Python program is fine.

There is no objectively "perfect" approach. I personally prefer getting a stack trace of the point at which the problem occurred. I find that more useful as opposed to getting back an error code which has been propagated and modified through layers of function calls and is often hard to correlate with the original error condition.