pgroves

Batching things into 'contexts' is the premise of Getting Things Done by David Allen. It was popular in the early 2000s. My guess is that the older engineer that told the author to batch things was referencing it [1]. It's a business book based on aggressively making and prioritizing todo lists and then working on them in contexts like 'email', 'phone calls', etc.

[1] https://en.wikipedia.org/wiki/Getting_Things_Done

When I read it on the ballot when I voted I immediately assumed it meant "libraries, etc", but now that you say that I'm worried "community areas" is bureaucrat-speak for "neighborhoods". Much different implications.

The same code split across two pull requests is preferable to the rest of the team (assuming they are actually trying to understand the PRs). Sort of a clever way to incentivize that.

Well then I'm glad you're here...Is it possible to integrate with an existing Dockerfile? My team's main build is handled by an image that is a result of years of fine-tuning and files with external dependency definitions (like requirements.txt for pip). Getting all of that installed in github's container is only slightly easier than getting it all installed in a desktop IDE, which is currently prohibitively difficult for us. We don't want to end up with a second platform we have to maintain.

More generally, is there going to be a clean way to to use this if we already have a containerized stack, especially editing code in a user's existing container? (With the code still under version control, of course)

It's blowing my mind that I just did a ctrl-F for "compound interest" in this thread and got zero hits. It's not even part of the discussion any more.

Treadmill desk. It's on the edge of being too difficult to walk and do real work, but you can walk pretty fast (for a decent amount of exercise) and watch videos easily.

Since you've made you're own app, and have noticed that a good card is a good tweet: I want an app that puts my cards in my twitter feed. I check twitter way too much and don't like opening Anki to review my cards. Not sure exactly how it would work, maybe I see a few tweets, and then there's a 'card' tweet, and I interact with it as usual and then get more tweets. Just an idea.

More time has passed between Office Space (1999) and today than between Office Space and Superman 3 (1983).

I use noScript with "Allow None by Default" on Firefox while I'm at work and should in theory only be browsing docs anyway. Trying to blanket ban third-party JS wouldn't really work b/c many sites have their legitimate JS on various CDNs. After enabling individual scripts on maybe a dozen sites I'm in good shape when it comes to web browsing day to day.

I wouldn't want this as a keyboard, I'd want one hanging from my right front pocket while looking at my phone with my left hand. It's the touchscreen interfaces that are terrible and need replacing, not the qwerty keyboard. The backlash against touchscreens has started with interior car controls and I'd like to see that spread everywhere else.

(Yes I have RSI problems.)

Is there a milestone the multi-threaded implementation could achieve that would make it a candidate to merge into core Redis?

Deskpass.com lets you drop into any coworking space in their network. It looks like they are up to 11 cities. You pay a subscription and then the coworking spaces get a flat fee for the day when you actually show up to their space.

Everyone on the team would benefit from strongly preferring short (single change) pull requests. Complicated PRs have a high cognitive load for anyone and this issue is highlighting that problem, not causing it. This is a good excuse to enforce a policy to package changes as small, separate PRs whenever possible. Although I might be reading too much into your comment about "significant amount of code" being the tipping point.

I generally think PR tools are all based on the same bad, diff-centric workflow that doesn't capture the narrative of the changes while bludgeoning the reviewer with minutia up front. The only defense my team has found is by making bite-sized PRs whenever humanly possible.

Sorry if I'm reading between the lines too much, but 28 cores and 56GB RAM for a business that is not covering it's own cloud costs sounds like a lot of compute resources. Unless maybe you're doing something unusual and have a lot of data when you don't have a lot of users yet. I would try to aggressively cut those numbers, not the price per core, and stay with DigitalOcean if it was me. I've been forced to use cheap, less reliable VMs and it's a huge time waster unless you have a big team and big dev budget to code around it.

You might need to slightly, "prematurely" optimize your computations to get by with less compute power, but if you're broke you're broke. Don't forget that the first few times you really need more compute resources, you can resize production to a bigger machine(s) in 10 minutes and that's not "wrong", even if it's not going to be the long term solution.

I've been on plenty of machine learning projects and have seen the boss make promises that the engineers are expected to keep. This is exactly what it looks like. The outcome is always overfitting a simulation or data set somewhere so that it looks like it's working, but the real world performance ends up being worse. Unless the engineers get stubborn and declare themselves to have failed (which can actually happen).

But in the projects I did, it was some dumb sales recommendation system. These people are doing it with self driving cars.

I see your wall of text, and raise you a wall of text:

I've started looking for a new job recently. I clicked the button on LinkedIn saying recruiters could contact me and have gotten over 100 incoming messages since Christmas. I'm not sure which buzzwords I have that are causing it. Many are junk, but not all. They all want a 'quick chat' to proceed. There are probably 20 I would talk to but getting started scheduling phone calls is going to be a huge pain given I have a software job that I do all day. Just responding to the 100 messages as "no thanks" or "let's schedule something" is going to take a day. I outsourced updating my resume and website, and am trying to get Hubspot working to set up a sales funnel (but that works horribly with linkedin email redirects, so not sure that's going to stick). I'm trying to get organized and would even pay for someone to actually represent me and help with logistics, but the 3rd party recruiting industry is built around pleasing companies for money, never the candidate.

So I'm not sure recruiters and hiring managers need more coddling and new tools to target candidates. They already have the candidates out-gunned by a few orders of magnitude. And they have resources (and their own time) to deal with their pipeline. I don't. Maybe they could spend a little time and effort differentiating themselves? Cuz these job ads all look the same, and they all think they're doing me a favor for allowing me to be considered for an interview. If they are getting a lot of resumes that look the same, maybe that's a signal about themselves and not about the overall population of developers looking for jobs?

And as for getting my past employers to provide recommendations... They are all hiring. It seems crass to ask them to write me a recommendation while making it clear I don't want my old job back. The time to ask for the recommendation would have been when I was working there.

I've been thinking about this lately. What if the web fundamentally was data collection of what a person is doing, and then things hang off that. Right now it is a collection of information that you can 'find'. Like your photos, web history, and fitbit would be generating your profile which is the fundamental part of the web, not the 'addressable places to get javascript apps' we have now.

Anyway that's as far as I got. Good luck and let me know how it goes.

Reminds me of how bond yields work. If the interest rate guess from 1% to 2% on an existing bond, the bond itself has to lose 50% of its value. I think the potato story is somehow less intuitive.

Based on the email from the VP, my wild guess is they figured out that Twitter operating Smyte would make Twitter itself not GPDR compliant (or some other big regulation), and they didn't figure it out until after the deal was done.

I don't see why your investors wouldn't be interested in pivoting the current company to a bigger opportunity with less competition. What are the objections going to be? They are not going to get mad at you for finding a bigger opportunity when you were out talking to customers.

....and if you're a recruiter, these are the issues to address in your sales-pitch. I can't believe I still get cold emails to work for 80hr weeks but only if I am good enough.

This sounds somewhat like "Privatizing Social Security" that was a big topic under George W Bush. The part of your taxes that go toward Social Security would instead go into something like an IRA that is restricted to only buy relatively safe things like diversified funds.

It's not quite the same but there is extensive existing material to cherry-pick ideas from. E.g. https://socialsecurity.procon.org/

I've used PocketCasts for quite a while now and would move away from it if there was a decent (cross-platform) alternative. It has some serious problems downloading (I think it actually hits their mirror of all podcasts, and the downloads will stall out for several minutes at a time). And for some reason streaming a podcast is the app's worst fear and it has a UX to do everything it can to stop you from doing it.

downdetector.com now seems to be down. never a good sign.

Predictive Model Markup Language (PMML) was released in 1997 to solve a very similar problem (but for lots of predictive models). I don't know that it ever really caught on. That was also promoted by the tool builders but not by the users.

While checking the wikipedia page, I'm surprised to see that it is still being developed (with a 2016 release): https://en.wikipedia.org/wiki/Predictive_Model_Markup_Langua...

Please note that in 1997 it was NOT ridiculous to use XML for this.

Illinois is one of the states where driver's license numbers are computed from all the other information: http://www.highprogrammer.com/alan/numbers/dl_us_shared.html

Best part is the woman who confronted the CEO about being a serial sexual-harrasser and ended up CEO.

Edit: Re-reading I can't tell what his title was, only that she was CEO after the fact.

More optimistically, it could be a step toward solving the problem if it adopted a Roku-like model where all those apps are installed and logged in at all times and you can pick one from a standardized menu. On Roku you can go to Netflix, Hulu, and single-channel apps like A&E.

Bonus points if there is a standard way to 'get a call' or 'search for a contact' that cuts across the video conf apps.

I know someone who is about to launch a very niche "box of the month" type service, where a box of items are mailed out each month. This would have been great as she could just have an amazon subscriptions page and a static website that links to it.

But no, it's for "digital app, website, or software" only.

Why would I go to amazon.com to sign up for dropbox? I'm not getting it.

I don't know Ansible very well... is it possible to install only a subset (namely, the vpn service you plan to actually use)? It seems like a large attack service to have every VPN software possible running on the machine.