Hacker News Reader

arochon TileDB closes $15M Series A for universal data engine

Hey Stavros. We were looking for a data-store to integrate into a clinical genomics LIMS that supports in-system analysis. We deal with de novo sequenced clinical samples (and not genotyped samples, which seems to be what TileDB-vcf had in mind?). There are some edge cases that TileDB-vcf explicitly disallows (updates/reinserts to the same sample, overlapping variants) that are not edge cases for us but rather common occurrences.

arochon TileDB closes $15M Series A for universal data engine

In genomics land its less an object store and more DB built to house giant but very sparse columnar data (think the position of genetic variants in hundreds of thousands of people) with support for interval operations and some other genomics things that traditional DBs don't support.

That said, we evaluated TileDB for genomics recently and found it lacking for our use case.

arochon Grant Imahara Has Died

Fair, I should have said "still on a ballistic trajectory". Given that a bullet fired not straight up may travel miles, I'm just guessing here, but legal was probably not too keen on them testing the more likely scenario. And, of course, every wants the payoff of full scale.

I totally agree that simulations of the most likely events would have been nice. The show could have benefited from more strict adherence to the scientific method. Alas, that probably would have made the show less appealing for a large segment of the audience.

arochon Grant Imahara Has Died

In fairness, if I remember correctly, the bullet myth was testing the narrow condition of "shot straight up into the air" such that the bullet would be returning at terminal velocity and no longer be on a ballistic trajectory. Whereas, most people struck by bullets fired into the air are probably being hit by bullets that are still on a predominantly horizontal trajectory.

arochon Portland approves 10% cap on fees that food delivery apps can charge restaurants

That's not what this is...? It's a cap on the commissions UberEats/Doordash/whatever can charge on an order. To, hopefully, prevent the commission structure from causing places to lose money on orders ($10 order with 20% Doordash; costs the restaurant $8.50; Doordash charges $2 commission; restaurant loses $.50 on the order)

arochon Engineering Dropbox Transfer: Making simple even simpler

If you use 7zip on Windows or just split/cat on macOS/Linux, you can split without compressing or archiving, saving you unnecessary CPU cycles (and space).

arochon The Waypoint NDA

Yes but they could edit it and not remove the Waypoint text like they're supposed to, and then you might be screwed

arochon Coronavirus: Dexamethasone proves first life-saving drug?

Given that it was discovered in the 1950's, I don't think you're going to see a patent on the versions currently being used in clinic. Now,if someone creates a more bio-available/active functionalized variant of it, fine that might get patented. But I doubt it would have much of an impact on the suggested COVID19 treatment (and would likely be years away from market)

arochon All-in-One DNS block list

They do? You can globally whitelist urls

arochon GitHub Classroom

Ha, I did something very similar also with Gitlab and runners. I was always amazed at the new and creative ways students managed to break the autograders we wrote -- particularly for our BASH scripting assignments.

I was also never brave enough to have my autograders submit grades directly to Canvas after the first time I did it. I confidently had it directly submit grades the first week during the first year I implemented autograders. Half the class forgot a shebang on their shell script, and the autograder happily gave them a 0.

For bonus fun times: Gitlab's runner timeout is pretty much nonexistent if you're using the shell runner. It just won't kill them if they take too long, so I had to write our own management system for runners. Silly university IT not support containers :(

arochon Show HN: I built a free forms service for static websites

Shouldn't your tag line be "The Form Backend Built For Developers", with "built" not "build". And a little more unsolicited feedback: the rendering/display of code example under /How It Works/ looks kind of sloppy -- I think it might actually look better when JS is disabled. Similarly, not a big big fan of the fullwidth Codepen embed.

arochon Ask HN: Did Google turn off IMAP access for good over the weekend?

Voice requires a phone number to link your GV number to

arochon GitHub is now free for teams

It's on the FAQ at the bottom of the announcement blog: https://help.github.com/en/github/getting-started-with-githu...

Though it does require a bit of between the line reading

arochon GitHub is now free for teams

The cynical thought would be drive usage of Github specific features/integrations to increase lock-in

arochon The first two years of my PhD

The vast majority of STEM PhDs are not staying in academia. The handful of recent grads I know who did stay in academia are, as I would put it, in academia-lite. They're in industry and privately sponsored positions or institutes, that pay quite well. Take a look at, say, the Broad Institute at MIT or the Allen Institutes in Seattle. Both are large, academic institutions that pay industry competitive wages.

arochon The first two years of my PhD

I don't think you have a great grasp on the PhD process in the US

A vanishingly small number of people are paying for their own PhD in the US. You are expected to be _funded_ though. Most programs give you 1-2 years of guaranteed funding (either through research or teaching assistantships) to give you time to find a PI if you didn't start the program with one and apply to grants. Having to secure grants is much different than paying your own way.

Most PhD students are either directly out of a bachelor's program or worked, maybe 2 years, in the field they want to do a PhD in. They're still most definitely in their early to mid-twenties. Are you thinking post-docs?

Length of stay is really dependent on your field of study and your specific work. I have friends who took 3 years to do computationally focused doctorates and some who took 6 years to do biology focused ones. You simply cannot make living things grow faster through sheer force of will.

I guess my last note would be that there's a reason US biology/biochem/bioME PhDs are paid a premium over their European colleagues internationally. And that's probably related to time spent getting their degrees and the depth and breadth of their experiences in the process

arochon Tell HN: GitHub Pages Down

How is Netlify's CLI spyware?

arochon Vagon – Personal high performance computer in the cloud

We generally use the browser interface, since many of our classes have both a GUI app and commandline work -- we've done this with both the Windows VMs and the linux ones. For linux desktop (Ubuntu), we use the ML-in-a-box image as our base and build on top of that. The linux VMs you can also SSH in via your favorite terminal/client

arochon Vagon – Personal high performance computer in the cloud

At work, we use Paperspace quite a bit for Desktop-as-a-Service in order to teach distance learning and in-person classes. We like not having to rely on our users having beefy(ish) laptops or being able to install software on them. The killer paperspace features for us, in no particular order, are:

1. Private networks -> keep classes isolated 2. Shared drives -> mount read-only class data 3. Create machine images -> build out the 'base' class image and spawn new instances from it 4. An API for spinning up and down

It seems like you are positioning yourselves as more for individuals and less for companies. It's not clear fo me from the landing page, but does Vagon require you download and install a client? Or is there a browser-only option? If we were to spin up, say 35 VMs, and then destroyed them a week later, is that $7.99/month charged for each machine or just once? Is it pro-rated if you don't use the whole month?

arochon Google's Abandoned Android Authenticator App

You've posted the same comment several times throughout this thread and in the past you've also, shall we say, name dropped saaspass repeatedly. While also ignoring questions about your affiliation with the company. Mind explaining why you bring them up so often?

arochon Launch HN: Probably Genetic (YC W19): At-home DNA testing for every gene

In fairness to 23andme, they also don't share your data without your consent ( Individual Data Sharing Consent : https://www.23andme.com/about/individual-data-consent/ and population-based Research Consent: https://www.23andme.com/about/consent/).

It just so happens that when they do share your data it may be with a Pharmaceutical company

arochon GitHub Enterprise is now free through Microsoft for Startups

You might also pay for stuff like LFS and packages.

But if I had to guess, $1000 is probably in the 95th+ percentile of spending for teams <20 (or any sufficiently small group) -> thus this is letting startups spend on Github resources without worrying about running out of credits for two years

arochon Accelerating Photosynthesis

https://rubyplants.com/ruby-team.html

Not linked on the current home page for whatever be the reason

arochon Google has handed information related to my account to the DoJ

Are you referring to the Steele dossier as "garbage 'research'"? If so, I don't quite understand your concept of garbage, given that much of it was proven correct. And I'd point out that the FBI says the dossier wasn't used for the FISA warrant.

Cue the deep state conspiracy theories, I suppose....

arochon Apple’s Activation Lock Will Make It Difficult to Refurbish Macs

They "recycle" them by going to Walmart and putting them into a recycling kiosk like [1] and get cash/gift cards instantly. They're not recycling so much as fencing....

[1]: https://us.ecoatm.com

arochon Firefox 71

There's a "Show All" that, well, shows all of the entries

arochon 16-inch MacBook Pro

You really don't though? The top of the line comprable configuration has a significantly worse display, 75% of the SSD storage (that's also significantly slower), a worse graphics card and costs ~$4700.

arochon What costs are exponentially dropping?

I mean, I suppose that depends on how you're defining "overhead" cost? To sequence a human genome, you _need_ to purchase a ~$700 reagent kit from Illumina to plop into your HiSeq and a ~$300 library prep kit -- are these overhead? I wouldn't call them overhead, and this fixed-cost hasn't really changed for 5+ years.

Those two costs are what makeup the $1000/human genome stat, we don't factor in the human and consumables components

arochon What costs are exponentially dropping?

Ehh, not really for any appreciable amount of sequencing, w've been pegged right around ~$1000/human genome for a while now [1]. The machine mentioned in the NYT article is almost certainly a variant of Thermo's RapidHIT system[2], which is not DNA sequencing.

[1]: https://www.genome.gov/about-genomics/fact-sheets/DNA-Sequen... [2]: https://assets.thermofisher.com/TFS-Assets/GSD/brochures/rap...

arochon Show HN: Disk.frame, a larger-than-RAM data manipulation package

Is this any different than Parquet (https://parquet.apache.org/ -- index, columnar data split across files), Feather (https://github.com/wesm/feather -- which has first class R support or FST (https://www.fstpackage.org/ -- Also first class R support)?