Danya was such an incredibly positive influence on the chess community, a tremendous teacher whose YouTube content I’m sure will remain popular for years to come, and my personal favorite chess commentator.
He was also only 29 years old.
I’m actually in tears right now struggling how to break this news to my son, who absolutely loved Danya and had a chance to play him OTB last year.
If you're using this on an iPhone, the failure is quite non-informative. Trying to browse to a URL looks like you failed to input anything and it just falls to a blank screen.
On the Mac, you'll see the message "Safari can’t open the page. The error is: “The operation couldn’t be completed. No such file or directory” (NSPOSIXErrorDomain:2)".
I think I have on 3-4 occasions played a game where, after evaluating on chess.com, got a 100% accuracy (which is engine correlation). A couple times were all theory and then blundering a mate in 1, but...
I did have one game where I didn't know the theory except a very vague recollection in the beginning. I actually thought I had blundered in that game and was trying to figure out what I'd do if my opponent made a certain move — they didn't find it, I ended up winning material in a tactic and they resigned — I was in complete shock when it came back 100% accuracy (and I definitely did not see the engine response to the move I was worried about, which was the best move).
I'm only around 1600-1700 on chess.com.
Not taking a position either way on Hans, but I have no doubt he knows far more theory than I do (and I do know some lines 20+ moves deep), and correlating with an engine is not impossible even outside of book.
To repeat what was said above, accuracy is not the same as engine correlation.
Engines often play moves that are counterintuitive and weird, but nonetheless good. This is because they can evaluate large trees of tactics in a way that humans cannot.
If a human finds a natural move that is just as good as the engine move (in terms of evaluation), they are still playing accurately, but they are uncorrelated with the engine. Playing accurately is not a sign of cheating. Playing many engine moves is a sign of cheating.
But how can you justify removing the ability to have a local vault?
Why would anyone think for a second that it would be a good idea to force people to store every password for everything in their life in your cloud without an opt out?
That, even more than Electron and the subscription model (both which do bother me), is an absolutely deal breaker. I've paid for every version of 1Password since v3 in 2009, but I'm done with it now.
The original article goes into great detail as to why we're moving away from local vaults.
That being said, we are looking into gauging user interest in self-hosting. Please take a look at our survey [1] if you want to share your thoughts. Hope that helps!
I don't want self-hosted, and I definitely don't want subscription based. I, 1Password user for many years, want the local vaults!
1Password 6 is great, and I'll keep using it until it quits working on my devices, but no more after that! I used to recommend 1Password so much to people it was borderline evangelizing, but I quit recommending it once the subscription was pushed over the other options, and now that local vaults are going away I'm actively recommending against it to anyone that asks.
Guess I'll be moving to Bitwarden or Keepass myself; time to research!
That looks a bit like NDJSON, assuming that the space between the objects is actually a newline.
There are libraries[1] that support parsing it and it's not too hard to do yourself, either. Some fairly popular projects use it to represent multiple responses in a single response body[2].
Both the websites are fancy ways of saying "one JSON object per line". The idea of laying out data that way was in common use before either of the website, they're just recording the practice.
We recently spent a couple weeks going over pricing with a GitLab regional Account Leader.
We already use GitLab for all our source repos, CI/CD pipelines, and had been using it for issue management as well — all fully self-hosted. The hope was to drop Jira, which was also being used for project planning, and fully adopt GitLab, but the costs were simply unjustifiable.
Basically GitLab ended up being a full order of magnitude more expensive. Even with discounts, which got things closer (but not all the way there), the fear was after a period of time, the discounts would be ended/phased out and we'd be stuck.
I'd love to see those features that compete directly with Jira (like roadmaps and multi-level epics) come down to the Premium level, which is more price/feature competitive.
We love GitLab, but find ourselves stuck using the free tier and paying for services we don't love, rather than supporting GitLab. I'd suspect we're not alone there, either.
Yea. Love gitlab but the pricing does add up quickly. Even if with all the features, it's really hard to get value out of Gold/Ultimate.
When I add up all the major SaaS apps an average user has: SSO, Email, Chat, Storage, Zoom, Support. Combined, GitLab Ultimate costs 2-3x that. And I promise you that as much as we all hate a messy inbox, email is more broadly useful than GitLab.
The other problem with ultimate is that it strongly incentives you to NOT let anyone else on the platform. At $1200/year there is no way in heck I'm letting the artists use Git, they can stick to their terribly Dropbox hacks. Marketing team working on assets with developers? Use email, no way you're getting access to GitLab. The $100/user/month model makes sense if they are a core developer (still too expensive, but makes sense) using every single feature in the system, but nothing else.
> The other problem with ultimate is that it strongly incentives you to NOT let anyone else on the platform.
This is my biggest problem as their pricing model discourages collaborative development.
We use GitLab to generate docs that are read by hundreds of internal users. On the free tier, if a user wants to suggest a change it’s no problem. Even though that is a very rare user and might only create one issue a year. Or maybe they add a tutorial or something to a project.
They aren’t developers, but having them involved in the git lifecycle is really helpful. Also data scientists who just want to archive their pipelines.
But with the ultimate tier those users suddenly cost $1200/year for minimal features. We can’t upgrade for free for the developers because we’ll disconnect all those “casual users.”
The suggestion to run two instances is stupid and confusing to users who now have to learn about mirroring, etc.
It’s weird that they don’t allow individual users to have tiers, we would buy more GitLab.
As of now, we will likely have to switch off of GitLab because there’s not a clear dividing line between software developers who need GitLab features and staff who write software who just need git, issue tracking, wikis and pages.
It's not a high number of accounts before you can have a person dedicated 50% of the time to just running a local gitlab setup with any options you want. Including infra costs + on-demand CI/CD.
Yes, I'm fairly certain it is. We switched to the SaaS because of that, it wasn't worth the hassle of self-hosting when you get all the responsibility and still pay the same price.
Yep. We upgraded from free to starter to get one feature. Now everyone is really careful about who to let on Gitlab because too many would bump us up to the next pricing bracket. We might even go back to free.
Ugg. I've seen that system used at other places and it's miserable. Devs had accounts but users and interns used a shared account. You get a vague ticket opened and they forgot to add their name. Also, would updates get blasted to everyone at the company? I'm pretty sure our's just turned them off for that account. So they don't get pinged when the ticket is resolved or needs feedback.
This is confusing to users. Seeing who made changes and who opened issues is really important. Having all users share an account would be confusing to see the same person asking and answering questions.
Also, making a thousand users learn a new userid and password is a support nightmare.
$1200 a year is expensive but you mentioned artists. Autodesk Maya is $1600 a year, 3DSMax is $1600 a year, Houdini is $500 a year, Zbrush is $900 a year, Adobe creative suite is $630 a year.
I'm not saying you want to add another $1200 a year to that. Just putting it in perspective to "artist" expenses
The argument here is that if you want an artist to have access to your instance - even if only to occasionally use features that are available in the free tier - you have to shell out the full $1200/year.
Something like Maya or 3DSMax is something an artist can reasonably be expected to use a lot of the functionality of the product pretty often. Expecting an artist to use the majority of a gitlab license on a daily basis is a bit of a stretch.
I don't use gitlab but github has issue and project planning. I'd expect every team member to use those. Further, github issues supports easy images and videos. Great for artists, and designers to track things. Assuming gitlab does the same I'd expect the same there.
Maybe there are better solutions and so you don't want your artists do that there.
It goes the same the other way. As a programmer I've always needed a license to the same 3d software the artists are using even though I don't use it daily I might need to write or debug an exporter or script.
In my unfortunate experience, having evangelized Gitlab in my company from the days when they made money through support and subscriptions, as expensive as Gitlab already is, it will only get more expensive a few years later.
It's nice that they keep adding these features, but in reality, we have already integrated most of the functionality they've added well before they get around to buying the company making the features they want to add.
So the actual practical effect on us is that we simply have to pay additional costs for those features that we don't use (or pay in time and money to migrate to those new features, with no real benefit and a massive downside of further increasing dependence).
We are are in a similar situation: We were using Github Team, paying I think $4 a month per account (between 10 and 15 users). I (head of the Eng group) love Gitlab and had used it for some time in personal/open-source projects. I managed to convince the team about Gitlab and we migrated, initially in the Free tier.
My first surprise was that Gitlab does not allow for Monthly payments...if I wanted to go into the Bronze tier, I would have had to pay a whole year in full. My startup doesn't do whole-year payments (quarterly or monthly) so that stopped me on my tracks.
I guess with the full weight of Microsoft, Github will out-price Gitlab. It's kind of sad because I prefer Gitlab CI/CD to Github actions (I just couldn't make sense of them).
Honestly curious what bad decisions you’ve encountered? I’ve used Gitlab CI/CD on ~20 or so projects of varying complexity (small to enterprise) and have found it enjoyable to use each time.
I'd be interesting in hearing more about some of these.
We're currently using GitLab's CI/CD for ~50 or so private repositories, covering ~7 different languages without any issues. That includes testing, Docker builds and documentation generation for most projects.
We only use private runners though, I don't know if any issues are related to their own runners.
So runners are a good example:
The runner config is stored in a config.toml on the runner. If you want autoscaling runners, it gets very complex so you might want to version control it or back it up, but you have to store your AWS secrets in this config.toml. It also contains hashes that have to be matched with settings in the UI, and of course, tags, which need to be tied into your .gitlab-ci. UI configuration changes take places instantly and there's no undo. So you have configuration in 3 places, only one of which is version controlled. And lets say you have some new runners with new tags -- you aren't going to be able to run an old pipeline on the new runner.
Talking about bad design decisions, I prefer GitLab's way to handle Pages over GitHub. It's just strange to store generated artifacts in a (`gh-pages`) branch. Instead, in GitLab it's stored as separate files and removed automatically after a certain period.
Agreed with other commenters, we use gitlab CI extensively, and we find it great. That's the main reason we use gitlab, the code/repository part is not that great and we don't use the other stuff
Same, we ended up pushing back our subscription for almost a year, in which we really could've used their service. Just because the upfront cost was so high. And when we finally got a subscription it was a silver subscription, which is great but doesn't let us drop Jira. The gold subscription cost is just insane. It might make sense if you're used to paying SV wages.
I also have a small dev team of 25 that wants to move to gitlab paid model but the upfront cost is too much. We need a monthly payment plan even if it costs more than annual one because I can't justify a high upfront cost to my ceo. We're rapidly expanding and expect to be 100 within 2021.
Many of these points come back to "it's easier or cheaper for gitlab this way" which isn't a great way to convince a customer. Some of which seem to actually be "due to our inability to onboard, manage and retain customers in an automated hands off way it's cheaper or easier this way" which is a terrible way to convince customers.
None of these is convincing, mostly because "its better for Gitlab".
I don't care. I want what's better for our company.
And right now, that is not paying for GitLab.
It sucks because I am an evangelist for GitLab, but I'm fighting a losing battle and we will be switching paid Github over free GitLab because of things like this.
This is not a new issue. There is an issue somewhere on the Gitlab repo that I’ve been subscribed to for more than two years. At this point it seems unlikely to change.
Variation of the same here. We did try to show, for example, that we could roll off some contractors if we shut down all of the various jenkins instances. But that wasn't enough to justify the "ultimate" tier, we went with "premium".
And the business case still had to include fluff to make it sound at all like a good idea. Note that it IS a good idea in this case, but for reasons that are hard to show as dollars.
It seems like they need something (low base + ala cart add ons?) to have a better pitch for Jira/Bitbucket self-hosted customers to switch.
One thing they could do is reduce the cost for non-developer seats. Lots of people need access to the issues, or builds/deploys, etc, without needing to commit code.
"One thing they could do is reduce the cost for non-developer seats. Lots of people need access to the issues, or builds/deploys, etc, without needing to commit code."
Exactly! Lower tier users would solve many of the issues.
Definitely not alone. We use the self-hosted GitLab that comes with MatterMost out of the box: great!
Now offer something to kill of Jira at a reasonable price and you've got another customer. The functionality is all there, it's just distributed awkwardly across tiers for many small outfits.
In theory we have this x,y coordinate system that shows the cost of a feature and the value of the feature. Which is great if you are living in a bubble by yourself, or some similar products. Making someone decide between CI and shared document writing is a terrible choice to have to make.
The other way to present this graph is the classic triangle, Cheap, Good, Fast, but the way we typically use that triangle suffers the same failure mode, because 'expensive' is some weird variant on cheap vs fast, as represented by Brook's Law.
I think the Discord Team has a different interpretation of this - the maintenance cost of the feature matters more to the user than the initial development cost. I can deliver you a feature now that will cost $X a month to operate, or I can deliver you that feature in a couple of months and it'll cost 1/10th that amount. Which means we can offer it to users at a price point they can afford.
It feels like Gitlab has, like so many of us, a bunch of features that create opportunity costs that the customers can't stomach.
The free tier users? I'm amazed actually at what the free tier offers to small startups. We have hundreds of gigabytes worth of docker images in the gitlab package registry that we can't even delete/cleanup properly (just keeping the last two would do for us), and Gitlab is keeping all that data for free.
The only price tier that is worth it is the silver. The gold/ultimate is just way too expensive. It's a shame that the jump is so crazy for a team of 10, from paying 3K a year to paying over 14K a year.
Not surprised. I rail on this all the time. One of my major problems is only the Ultimate allows free guest users.
Its a big reason we haven't paid for GitLab. And we are also looking to go toward Github for that reason and some others too (and those others might be solved if we paid for support, but see the first reason).
I use a couple dozen tools on a regular basis. JIRA is by far the worst. The functionality is ok. Maybe it is a bit bloated. Maybe our process is a little over complicated. That's not a big deal. The thing that really gets me is performance.
The performance is absolute dog shit. Every interaction with it painfully slow. For example, the page to view a single issue is almost 20MB fetched over 100 HTTP requests and takes 10 seconds to load. This is without cache. With assets cached, it is still 4 seconds to render. This is the fastest interaction by the way. Everything else is worse.
Maybe some of this is my organization's fault. I really don't know. What I do know is that it is so slow that I dread every interaction with it.
Nope. We _just_ switched to Jira/Confluence from Notion+Trello (sigh), so we really don't have a complicated set up, and it is definitely "absolute dog shit" for performance. It is hands down the slowest tool I use. I've used it at large organizations with hundreds of devs and it's even worse. I regret not pushing back harder, since it is totally overkill for 1.5 PMs managing 3.5 devs.
Seconding the performance issues. This is obviously amplified when screen sharing, which makes certain types of collaboration difficult. (Remote grooming sessions, for instance.) I also find recent UI changes have made the product harder to use, at least for our version and processes, and have added nothing of value in return.
As far as how useful Jira is, a huge percentage of that is going to be your team's processes external to Jira. I have found it to be a very good mirror. If your team's processes are a mess, your Jira instance will quickly become one as well. But if your team is organized and is bringing an existing, functional system to Jira, I have found that it's a good enough value add to be worth it.
It's definitely an enabler, though, rather than a tool that will funnel you into good practices. (The same can be said for Confluence but it's even more true there IMO.)
Same performance problems here. I hate it. I also need to access other JIRA boards with different email addresses occasionally. It takes forever to sign out, login again, navigate to board, find the issue. It puts me off using it so much.
One light at the end of the tunnel is the Mac app (for some reason there is no Windows or Linux version). It's much more lightweight and feels way faster than the web version. However, it is buggy in my experience, and a fair few things I do a lot can't be done in it and require the web interface, but for quickly checking stuff it mostly works.
Hi - I'm coming from the Confluence team working on Performance, but I can say that we (all of Atlassian) are aware of performance is an area needing improvement, and we're working on it. Customer feedback (including logs, recordings, and exports of slow pages) is always very helpful to us. If you're open for followup over email to provide specific information, please let me know.
It's what you make of it. Most companies make it into a mess. And the UI is slow and bloated like a dead whale.
It is however not that expensive which matters especially when most of the company needs accounts and not just engineers. $7/user/month for basic or $14 for everything you reasonably care about. Gitlab is $19 or $99/user/month.
Jira's performance is absolute garbage. It's painful and embarrassing.
I'd use GitHub if my team were willing to do non-technical project management with it as well. I wish GitHub would add features for non-technical users on projects. It would be a huge win.
Literally just spend some time on some docs features and provide a view of projects that isn't centered on code for those users that aren't involved in the code.
For a reasonably technical user, you can make it work, except that the views just aren't built for them, so it's a lot of visual clutter that isn't necessary and holds the product back.
Been looking at Clickup since, like Jira+Confluence, you can integrate wiki/web documents with your project management tooling. I've found that this is a pretty critical feature. I've tried repeatedly to get folks to rally around Google Drive or a folder of Office documents, and it's just not the same as working in a single, integrated web experience.
In my case, I´ve used JIRA for a long time (8 years) and consider myself a "power user" (I've connected Lambda's with its API and webhooks, later used their "actions" scripting, I've delved deep into workflow modification, projects, board, views the whole gamut). It is a very powerful tool that if you REALLY know how to use it (like, really make sense of permissions, workflows, hooks and other features) you get really good ROI.
Nevertheless... it is an absolute and horrible hog. It is so slow and clumsy that it is frustrating. Also, when they changed the interface they kind of hid a lot of stuff that used to be there. Also the free "Gantt" feature they have sucks, and the paid one is OK but does not justify upgrading 50+ users to the plan that makes it available.
It's almost expected you don't have enough time. When I was at West Point they'd explicitly state that.
Additionally, there was a hard lights off policy at midnight to ensure cadets got enough sleep, yet people invariably would be up with a flashlight underneath their covers trying to get work done until 2am.
Plebes in particular got less time as well, as they'd be subject to additional tasks/hazing until a given time when they couldn't be disturbed (I believe 7pm).
It's important to remember that almost every cadet here was a plebe (1st year student). Plebes are under enormous pressure, and they are only just beginning the inculcation of the honor code (they've all only been there about 6 months now).
Traditionally, they are given by far the most leeway. The expectation is that going through this process will only hasten and harden their understanding of the importance of the honor code.
I recall reading stories in the news about happenings when I was cadet, and there's always invariably some nuances and details that are either incorrect, missing, or misunderstood by people not familiar with West Point.
Some cadets are also still going to go before an honor board, which may very well lead to their separation.
Without knowing more details, I trust the administration (including cadet-led) to take actions in the best interest of our country, the institution, and the Corps.
Ironically, that knowledge leads to a crisis of trust in military institutions among enlisteds.
The trust problem is critical and far ranging, but also multifaceted. It permeates not only the military but all of US society at this point. I'm honestly unsure as to the best solutions? We may only be able to restore trust with glacial change rather than sea change.
He was also only 29 years old.
I’m actually in tears right now struggling how to break this news to my son, who absolutely loved Danya and had a chance to play him OTB last year.