It will be true because of the "within a visibility timeout" right? Of course that makes the claim way less interesting.
I took a peek at the code and it looks like their visibility timeout is pretty much a lock on a message. So it's not exactly once for any meaningful definition, but it does prevent the same message from being consumed multiple times within the visibility timeout.
> it does prevent the same message from being consumed multiple times within the visibility timeout.
... When there is no failure of the underlying system. The expectation of any queue is that messages are only delivered once. But that's not what's interesting: what matters is what happens when there's a system failure: either the message gets delivered more than once, the message gets delivered zero times, or a little of column A and a little of column B (which is the worst of both worlds and is a bug). If you have one queue node, it can fail and lose your data. If you have multiple queue nodes, you can have a network partition. In all cases, it's possible to not know whether the message was processed or not at some point
The Two Generals Problem is a great thought experiment that describes how distributed consensus is impossible when communication between nodes has the possibility of failing.
Distributed consensus within a quorum is possible. With only two nodes there is no quorum. With three nodes and at most one offline, consensus is possible.
YJIT and Ruby 3.3 have really impressed me as well. Their VM engineers are clearly doing something right.
Related to Ruby perf, I still hear folks worried about rails “not being able to scale”. Let me say something controversial (and clearly wrong): Rails is the only framework that has proven it _can_ scale. GitHub, Shopify, AirBnb, Stripe all use rails and have scaled successfully. Very few other frameworks have that track record.
There’s plenty of reasons to not use rails, but scaling issues doesn’t feel like a strong one to me.
But, for the sake of truth:
- AirBnB migrated from Rails to a micro-services architecture (which I think, they regretted doing too early - I read that somewhere I believe)
- Stripe never used Rails: they use Ruby (and Sinatra for the Web part - i.e. dashboard).
But it's true that Github and Shopify both use and scaled Rails monoliths. There are showing the way :)
To provide some context for this, Atlassian originally created a drag and drop library called react-beautiful-dnd. A few years ago they stopped maintaining the library, and some community forks took over (I think @hello-pangea/dnd is the most popular).
In the interim I've found https://dndkit.com a _really_ nice solution, but it's only available in React-land.
Looks like this new library was written at least in part by the same guy who wrote react-beautiful-dnd, so my hopes are high (as long as this one stays maintained...)
Both the dndkit and pragmatic-drag-and-drop are great for simple use cases. If you're building something complex though, neither will cut it, nor will any other dnd library I've seen. I built a card game interface last year that involved 3 forms of droppable:
- a hand, where cards were display in a row and could be reordered
- a pile, where cards stacked and only showed the topmost card, which could be dragged off the pile
- a play area, where cards were freely arranged and overlapped each other based on which was moved most recently (and thus "on top"), and in which groups of cards could be moved as a unit via drag-select
For cases like this, these libraries prove woefully inadequate. I had to hand-write the whole thing and it was by far the most challenging front-end work I've ever done, especially since it's multiplayer.
To be fair, this is like saying Legos aren't appropriate for building a house or a pocket calculator isn't appropriate for analyzing a particles accelerator -- simple tools are great for their intended use case but you need much more advanced tools for complex scenarios.
With that said, I agree that emulating playing cards is tricky, I'm curious what you used. I built https://play.cheatatbj.com using ZimJS (whose web site looks kind of like a joke but is actually a very impressive animation framework).
I didn't use anything for DND, I wrote that all by hand. I used React, Pullstate for a state store, and PubSubJS. All of them are great to work with, although this was really, really pushing the limits of what React is made to accomodate. If you play Magic: The Gathering you can check it out by starting a mirror-match by yourself at "ca rd tav ern dot com" with the spaces removed (don't want this comment showing up in search results), but it would require you to make a deck since I haven't gotten around to adding default starter decks to new accounts yet. (The site was largely finished but never launched or shared.)
What's the point of this? To get better at blackjack?
I ask because I did a blackjack trainer to train me via repetition on perfect play (circa 2008). Still one or two things left unimplemented to this day, though, but over the course of playing a few games a day over months, I definitely did get better (losing less over time than I used to :-))
Yes. My UI's functionality was what I would personally describe as "unbelievably complicated," but I don't know how I would have done it without the modularity of React. It's a full-blown sandbox for playing Magic: The Gathering. I'm sure it could be done otherwise, but React is what I know and it didn't make sense to me to work in something I didn't know rather than something I knew well.
I’m just wrapping up a project at work with similar constraints and share the same experiences. Good library, best used for simpler use cases. Drag and drop across browsers and devices is still not a smooth dev experience. Dnd did improve it a lot and performance issues are not noticeable to the end user in our use case. Hopefully the new version that comes out continues to be as good.
I've got my hands pretty full with another project and unfortunately I don't use Twitter anymore since it became X. If you'd like, though, I can send an email to your quilt store's contact email (team@) and you can reply to that to tell me a bit more about what you're trying to do, and if I can give any genuinely useful suggestions as to how I would go about it, I will do so. But I realize technical advice != actual developer hours and if that's not so helpful, no worries.
If something(s) changed there would you have stayed or would you return?
By measurable account he's made the platform better overall, contrary to certain propaganda narrative talking points.
He did of course make Twitter-X less of a bubble machine, which arguably is bad for society because then ideological mobs with certain unchallenged ideas or beliefs go unchallenged (doesn't matter what "side" you're on, there are extreme and bad-wrong ideas on both/all) - so you'd begin to get exposed to content that may make you uncomfortable. If you're not able to quickly scan through it and not react so you just move on, it would become a tiresome and a challenge.
You could go "full bubble" like Instagram recently did though and create a setting default on to not show you "political" content from people you don't follow - arguably adding to the detriment of society, and of a seeming concerted effort by bad actors towards division and conquering us by helping reinforce the bubble walls.
I find your perspective fascinating but I never ‘got’ Twitter. My current perspective is that a bunch of idiots are loudly shouting the shortest/most successful piece of mimetic garbage has currently evolved and the day I see something useful arise out of that morass of RNA-analogue I’ll be amazed but until then I wish it would stop taking attention away from larger more nuanced conversations which are more than a couple hundred base pairs (or tokens) long that we need to be having in order to navigate our current waters
dndkit seemed very promising, until I realized the state of development [0], and critical performance issues if you want to use it for larger lists (individual issues linked within [0]). So it's fine if you need to use it for simple scenarios and a smaller number of items, but otherwise, it's inadequate in current state, unfortunately...The fact that the post was in Aug 2023, perf. issues still remain and there's low activity on issues in general, doesn't bode well...
I also found dndkit and built a very complex/robust drag and drop form builder with it. It gave me nice abstractions for custom collision detection without worrying about the browser implementation. Highly recommended.
This seems incredibly important. I know non-compete rules personally held me back at a previous tech job.
I'm interested to see how this hits finance firms – I know people who were forced to take a year off between jobs (although they were compensated the whole time). Always thought that would be a pretty sweet deal.
Page 83-84 provides some guidance on garden leave and suggests that it will still be allowed under the new rule:
> With respect to garden leave agreements, as noted previously, commenters used the term “garden leave” to refer to a wide variety of agreements. The Commission declines to opine on how the definition of non-compete clause in § 910.1 would apply in every potential factual scenario. However, the Commission notes that an agreement whereby the worker is still employed and receiving the same total annual compensation and benefits on a pro rata basis would not be a non-compete clause under the definition, because such an agreement is not a post-employment restriction. Instead, the worker continues to be employed, even though the worker’s job duties or access to colleagues or the workplace may be significantly or entirely curtailed. Furthermore, where a worker does not meet a condition to earn a particular aspect of their expected compensation, like a prerequisite for a bonus, the Commission would still consider the arrangement “garden leave” that is not a non-compete clause under this final rule
even if the employer did not pay the bonus or other expected compensation. Similarly, a severance agreement that imposes no restrictions on where the worker may work following the employment associated with the severance agreement is not a non-compete clause under § 910.1, because it does not impose a post-employment restriction.
Doesn't this mean the end of 'at-will' for anyone that a company wants to cover with a non-compete? At present a company can have their cake and eat it, giving you one day notice of end of employment but then enforcing a non-compete for x months.
If this survives the Supreme Court, wouldn't a company would have to put in your employment contract that you/they must give x months notice to end employment if they wanted to restrict you? Otherwise you could give one days notice and they wouldn't be able to put you on gardening leave.
The gardening leave still works as long as it's beneficial to both parties. The employee gets an effective long, paid holiday and the company gets a non-complete equivalent. Now they have to really ensure that pay is worth more than changing the employer though.
It's less "end of at-will" and more "if you want effective non-competes, it's going to cost you".
Or the employee agreement you sign includes a notice period of 6 months from either side.
I think you're right that negotiated gardening leave will happen but I can also see companies baking in the gardening leave at the start of employment (so no longer at will) as being cheaper.
Source? My understanding is that at-will states have a default presumption of at-will employment in the absence of a contract, but parties are free to contract alternatively. Which states invalidate mutually agreed upon notice periods?
It’s not illegal, but from what I understand an excessively long notice period would very likely be shot down by the courts. There are a lot of arguments to be made here (for example, what’s stopping an employer from putting a 5 year notice period from day 1 if they can put an arbitrary one?). I imagine if companies do increase it or introduce one in general, it would be closer to 1 month, maybe 2, not 6.
They are not illegal, and even if there is a grey area good luck defending.
I worked at a hedge fund in NYC. I gave my 2 week notice. HR said that the employee contract that I signed when I joined stated a minimum of a 4 month notice period. This would have completely killed my next job which I already accepted.
I should have known this was in the employee agreement, but I didn't know they they would threaten lawyers on me if I didn't stay 3 months.
I was denied a job I was well qualified for because, (paraphrased, besides the quoted part): Our CEO and your CEO have a "gentleman's agreement" not to hire people that work at eachother's company.
I have no idea why the recruiter was willing to put this in writing, and thankfully, I was able to find other work instead.
I know it's not a non-compete, but there are other ways that companies can illegally form cartels to suppress labor.
I'm not a lawyer, so I wouldn't make any money. I know it's a common refrain, but look at what happened with the Google & Apple class action. Those folks got less than $6,000 each for what must have been hundreds of thousands in damages each.
It's also career limiting to be the person that sued their employer. I didn't have any monetary damages, as I was able to find alternatives.
For the record, in this context your payout would come from the sweet, sweet pot of funds the lawyers divy up. And it'd be a lot more than your direct monetary damages as the lawsuit would likely impair your ability to later work in the industry.
The "class actions never make money for the plaintiffs" refrain is 99.8% true. The people who get a postcard mailer about their eligibility usually get shit. The 0.2% outlier are the people who put themselves at risk by providing the testimony that wins the case, and they're rewarded for it. And the blood-sucking lawyers of course.
It got them a slap on the wrist and a nothing fine for years of depressing wages. No one went to jail and the head of HR at Google during this period also got a book deal.
I was subject to one as well, but it was just on my base, not total comp, which was not the majority of my compensation. And while my base was fine, it was more or less explicitly stated that this was meant to make it painful for employees to leave and had almost zero to do with any special information the employee had.
They have other tricks. My comp was about 80% bonus, most of which went into deferred comp for a few years. If I was to go to a competitor without permission (independent of the non compete) I would forfeit the deferred comp.
But at a certain point that deferred comp is “enough” that if you just go to work each day and hide in the loos it’s worth waiting around and collecting the cash. And the company won’t benefit from a mostly checked out workforce
The bar is quite high at these kinds of companies. If you let off the gas and try to coast like that, you'll just get fired and thus lose the deferred comp that way.
Reject the handcuffs. Once enough people do, they will stop making it deferred. They know that if you're willing to reject the money, you're willing to leave (the handcuffs aren't keeping you there), and that scares them.
I'm not sure that I understand. What does it mean for a noncompete to apply only to base compensation? Is the idea that if you join a competing company within X months of leaving your old company, you need to repay your base salary to the old company?
I think they are saying that the compensation they their previous employer paid them to not work for the competition for a year was based on their salary, not their salary plus bonuses, so it was not as good a deal as it sounds.
It means that GP was paid their salary for the non-compete time. In finance it is common for total compensation to be the salary plus a "bonus" of 100% of salary in normal years + any performance bonus. This means that if you had a non-compete in the finance industry and you left your employer for a competitor, then your previous employer could pay you your salary (meaning 50% of your usual compensation) to not work for that competitor for a year.
(These numbers are typical of finance industry compensation and non-compete terms.)
No, it just means that during the period after you stop working at the old job but before you can start working at the new job, you are paid only your base. This can be a significant reduction in total comp in industries such as finance.
I have worked jobs where the best bonus over 5 years was $500, while the typical year all we got was a promise that if things go well there will be a bonus. I've worked other jobs where the worst bonus was $15000 (a really bad year for the company), and could be up to $50,000. This is as a regular engineer, management can get a lot more. The first company taught me at until the money is in my account the bonus is meaningless. The second taught me that they aren't just a rumor. Most companies don't even pretend to offer a bonus which is acceptable - at least I know what I will make.
I think everyone should make 2-3x the poverty level income (we can debate exact numbers), and everything after that is bonus. So long as the company pays a bonus most years it means in a bad year you have enough to live on and don't need to find a new job, while in a good year you have a nice bonus to buy nice toys.
In finance, it's common to see a base of 150,000 and a VR of between 300k and 750k for engineers. During garden leave, you get paid your $150,000 as part of payroll, but are ineligble for VR. Your total comp goes from 450,000 (in mediocre years) or 900,000 (in good years) to 150,000 for whatever your non-compete period is (6 months, 12 months, 24 months are all common).
In finance you make more money, but work more hours and have more stress per hour typically. Still probably a good deal, but it's not a Pareto efficient deal.
>I know people who were forced to take a year off between jobs (although they were compensated the whole time). Always thought that would be a pretty sweet deal.
If a company wants to pay someone not to work for a year, they're free to do that whenever they want. Maybe without noncompetes, they'll have to pay more to make it worth it for the guy being paid to sit around!
We'll have to see what the finance industry does. My guess is that they will only make sizeable counter-offers to key employees, and the employees will not be forced to accept them and typically won't. In the long run there might not even be sizeable counter-offers to key employees.
EDIT: Er, the FTC explicitly does not comment on garden leave:
> With respect to garden leave agreements, as noted previously, commenters used the term “garden leave” to refer to a wide variety of agreements. The Commission declines to opine on how the definition of non-compete clause in § 910.1 would apply in every potential factual scenario. However, the Commission notes that an agreement whereby the worker is still employed and receiving the same total annual compensation and benefits on a pro rata basis would not be a non-compete clause under the definition,350 because such an agreement is not a post-employment restriction. Instead, the worker continues to be employed, even though the worker’s job duties or access to colleagues or the workplace may be significantly or entirely
curtailed. Furthermore, where a worker does not meet a condition to earn a particular aspect of their expected compensation, like a prerequisite for a bonus, the Commission would still consider the arrangement “garden leave” that is not a non-compete clause under this final rule even if the employer did not pay the bonus or other expected compensation. Similarly, a severance agreement that imposes no restrictions on where the worker may work following the employment associated with the severance agreement is not a non-compete clause under § 910.1, because it does not impose a post-employment restriction.
My guess is that garden leave will be offered, but in right-to-work states there will be no way to enforce that the employee remains employed.
I am currently on one of those deals by working for an HFT, then taking a competitor's offer. It is really very nice. From a wealth-accumulation POV, I am losing out a lot of earning each month I'm not working, but I am still paid a very cushy six-figure salary that covers a comfortable lifestyle for my family plus decent savings. I value my time at prime working age much more than the net worth I potentially lost. I have been able to travel, hone hobbies, start and finish personal projects, just help out my wife, and much more. Honestly I don't want it to end.
Since you are still employed by your previous firm during garden leave you are still covered by your employer’s health insurance policy. I’m also on garden leave at the moment and that’s how it works in my case.
COBRA just means that you’re continuing whatever plan you had under your employer, at full cost to you (vs. the employer presumably subsidizing some of the cost while you’re employed by them). The amount you pay is entirely dependent on the plan(s) your employer has chosen to provide its employees.
For me, the full cost of my current employer’s health plan is about $1800/month, and a comparable plan from the healthcare exchange is $2200-3000, depending on the plan. So if I were to lose my job, it’d be significantly cheaper for me to use COBRA than get a plan from the exchange.
It's worth noting that the so-called "garden leave" you're describing usually doesn't come with things like bonuses. That may even be a majority of your compensation depending on the role.
> The vote on the final rule, which fell along party lines, with three Democratic commissioners voting in favor and the agency’s two Republicans voting against
Is this some kind of weird attempt to ask me to teach you how to do research? You've got the whole Internet in front of you, same as me. Go do your own research and come back when you have a point to make.
I have yet to see a Republican led government advance banning of non competes (or really anything that benefits W-2 workers as a whole) in the last 25 years.
Washington (Democrat led), I think, most recently passed a non compete ban for those under a certain salary, but I cannot think of any Republican led states that have advanced such legislation, or espoused views that they want to.
It falls in line with similar worker friendly legislation passed by Democrat led states such as longer family leave, paid sick and family leave, higher unemployment benefits, higher minimum wages and minimum salaries for exempt workers, eliminating non tipped minimum wages, and publishing of salary ranges on job listings.
Edit to respond to below:
Is it partisan in California? If anything, I would have thought the California non compete ban is the most un-partisan issue since it has been in place since 1872, so neither of today's parties would be credited with it.
Another interpretation is that the ruling party is bribing people now that election season is ramping up by passing rules it knows has no standing in court, but won’t get shot down until post-election. Imagine all the people who voted for Biden thinking he would absolve them of the contract they willfully entered to pay their student loans. This is not much different. It is another group of people who have contracts they wish they didn’t have relying on government overreach to save them rather than not having put themselves in the position to begin with.
I've heard this talking point before, and it's really silly to frame it as shenanigans somehow because of the timing, rather than elected officials enacting popular policies that the voters want.
There is a body of actually elected officials whose job is to create legislation to enact public policies. The FTC is neither elected nor capable of creating legislation. This is an overreach of their power and will likely be ruled as such in court.
The same thing happened with Roe; it is not the job of the Supreme Court to enact public policy. That is for the legislature to do. If you talk to anyone in law, they will tell you that, regardless of their personal opinion on abortion, Roe is perceived as one of the worst court rulings ever because it was specifically designed to “legislate through the court”. Here we are in the same position again, with a non-legislative body enacting public policy; we don’t learn.
"Under this Act, as amended, the Commission is empowered, among other things, to (a) prevent unfair methods of competition and unfair or deceptive acts or practices in or affecting commerce; (b) seek monetary redress and other relief for conduct injurious to consumers; (c) prescribe rules defining with specificity acts or practices that are unfair or deceptive, and establishing requirements designed to prevent such acts or practices; (d) gather and compile information and conduct investigations relating to the organization, business, practices, and management of entities engaged in commerce; and (e) make reports and legislative recommendations to Congress and the public."
Noncompete acts are an unfair method of competition (see (a)), which Congress granted the FTC the power to define and restrict (see (c)). It is, quite literally, their job.
I was surprised today to find that Spectrum doesn't enable IPv6 by default. Only figured this out after way too long trying to ssh into my IPv6-only Hetzner instance.
Ended up just paying for an IPv4 address to avoid the hassle.
I don't think this is "the year of IPv6" but it might be its decade.
Might be a regional thing or something to do with your router. I have Spectrum and my router gets a global IPv4 address and IPv6 prefix over DHCP(4/6) just fine without doing anything special.
I have Spectrum at my house & while they will give me an IPv6 address it doesn't really work. The network offers up a wide range of IPv6 blocks I can choose to use. But whatever IP you use first, that is the only address you ever get traffic for. So the delegation is basically just 1 IP in practice.
Also IPv6 doesn't actually work reliably. You'll notice it when browsing the web for example, there are periods of 2-3 seconds without IPv6 connectivity. This does not happen on IPv4.
I was dealing with this the other day. Spectrum was giving me IPv6 - annoying when developing an IPv4 service. Ended up switching router to one that only support IPv4 and it reverted back to IPv4. If I plug my MacBook directly into the modem, get ipv6.
>I'm hoping/assuming/guessing those won't have that odd IPv4/IPv6 switching behaviour...
As a Spectrum Business customer, I wish I could tell you whether or not that was true, but IPv6 isn't available to me at all. I'd happily (okay, not happily, but I'd be willing to do so) pay for an IPv6 block (I currently pay for five static IPv4 addresses -- one of which is eaten by Spectrum's required router, the other eaten by my own router, but that's a different discussion..Grrr!), but I can't even do that -- as it's not even on offer here (NYC).
It seems that there's wide disparity in how/where Spectrum implements its IPv6, so YMMV.
I have used Spectrum and they gave me a v6 route without asking. Maybe your DHCP client or OS was to blame? I am not sure if it shook out via router advertisements or DHCPv6, it just worked with dhcpcd.
SchemaStore is a collection of JSON schema files. It's been down for about 9 hours now. You may lose autocomplete on json files in some editors because of this.
July automates the brand deal/sponsorship process for content creators so they can focus on creating great content.
We just closed our seed and are looking to bring on talented engineers in our NYC office.
Our stack is Typescript, Next, Prisma, Postgres, Redis, all managed with AWS CDK. We care deeply about quality and preserving type safety across our stack.
I’m making a permissively licensed json/whatever encoding you’d like English dictionary by parsing Wiktionary. There’s a beta demo type thing at https://www.mostaccioli.club if you’re interested.
CGP Grey made a great video [0] pretty similar to this, it has a bunch of points about physical things (changing sleep schedule, staying indoors) and keeps the same sarcastic tone.
I took a peek at the code and it looks like their visibility timeout is pretty much a lock on a message. So it's not exactly once for any meaningful definition, but it does prevent the same message from being consumed multiple times within the visibility timeout.