How does this platform still have developers

As far as I can see, this change mostly just gave people an opportunity to wear black face for racist jokes in the company chat without getting fired.

@Lensflare It wouldn't have, this is just plain a shit idea. What I'm trying to say is that if a change, any change, actually has a positive impact, however small, that would overrule convenience, and if it has a negative impact, convenience isn't necessary to defeeat it.

@Lensflare I think some social issues, and race in the US seems like such an issue, are important enough that if this actually helped anyone it would justify that bit of inconvenience. Where I live the racial tensions involve Gypsies, Arabs and East-Asians, none of whom match an identifiable emoji skintone, but I certainly think that convenience is a lame argument given the stakes.

There is one callback actually, sendRequest passes processResponse to fetch. processResponse then decides which request must've been sent and what to do next based on boolean flags in the component state.

Yesterday I had some business with a KnockoutJS viewmodel that didn't use any asynchrony, callbacks, nor lexically captured anything besides the viewmodel instance. Instead, each event had an associated member function and those all decided what to do based on a bunch of boolean flags and enums on the component state. It was the kind of event handling logic you would only be forced to write in C nowadays.

"Has anyone used Zustand? Is it a safe bet for the new timesheet portal or should we stick to Redux?"

👍🏼12 👍🏽8👍🏾1

Congrats, your conversation is now Racially Charged! ⚡

My coworkers all use the skintone modifiers for comedic purposes (ironically, they're the perfect device for racist jokes) but I imagine if they're used correctly it really has to be surreal to have racial distribution statistics on every single poll you ask within a team.

@Hazarth They must never have seen a chinese person for that colour to remind them of chinese. Like that association can only happen if "chinese have yellow skin" exists as a sentence and not an image in one's head.

🤝🤝🏻🤝🏼🤝🏽🤝🏾🤝🏿 No mixed race handshakes. Is the Unicode consortium promoting segregation?

We play music in the office on Christmas and a few other select holidays, and when there's very few people in and they agree on something. Otherwise hell no.

Hash functions are rarely fully reversible, usually a weak hash will either reveal partial information about the input or reveal only some inputs. Technically, if you have a massive lookup table, a perfect hash is a perfect compression algo.

Technically a hash that is reversible on 20% of inputs is a lossy compression algo; you can recover 20% of the input in a statistical sense.

The formatting looks goofy, before I read the rant, I was like "'wow, can pgp keys be this short?"

no, it would melt.
Now if you put it in a second, bigger freezer...

inspiring

@tamagotchi Some people keep an archive of their favourite XKCDs. Some remember keywords; most of XKCD is indexed. At the very least, explainxkcd.com has longform text that will contain any relevant keyword you can remember.

@tamagotchi I have to believe via AI given the pace.

@tamagotchi Hetzner decided (in 30 minutes including the appeal and rejection) that I'm not similar enough to my passport photo, so I'm not allowed to do business with them.

@jestdotty Well a small business owner whose idea of a successful life is selling lots of cookies and not spearheading the fight for independence would probably do whatever the powers that be demand from it, because making a powerful enemy is probably worse. Same with the customers, sure some would rather never eat a cookie again, but ultimately most people buy from whoever sells what they want where they are. That's my point, you can't engineer your way out of centralised power, the violence monopolized by the state can be invested into controlling anything.

@jestdotty yeah, but that's exactly my point; your activity on the blockchain can be connected to you in many ways, of which a nation state's preferred method would probably be to pressure the fiat exchange and to track known small users such as small businesses that accespt crypto.

Actually, in a sense crypto is less anonymous because everyone can access the transaction log and trace back the trades, whereas in traditional banking only few parties are. I don't have nation state level enemies, but I've known a few creeps, and I'd much rather know that they can't access my transaction history. The same is true for most of society.

@jestdotty I was talking about transaction data. How they each identify a user varies, but even in the best case a crypto user is only as anonymous as their trades are arbitrary and their fiat exchanges inaccessible.

I use Contabo, it's pretty cheap because it has comparatively low availability (99%)

@jestdotty The data is WHY we use banks. Their primary purpose is to accept cash, track its ownership status, and return it on-demand. They can't not have this data. Even crypto has this data, it just has a very flimsy notion of identity. Cryptographic identity is difficult for third parties to follow through to other services, and is completely unsuitable for things like account recovery - a pretty important feature of banks given how well the average customer handles their identification methods.

Selfies have fuck all to do with this, your national tax authority has access to your bank records without explicit notification. I swear, people have such shit intuition for privacy. The state has easier access to your transaction details including the ID of everyone you sent money to than an up-to-date photo of you because they need to fish the latter out of CCTV via all sorts of expensive ML tricks.

If you're forced to attach a selfie to money transfers, the selfie is the unusual detail in that file, not any connection they could make.

I guess there's also data about the geolocation of the transfer. That should be new too.

@12bitfloat You would usually only call the wide algo on the well-aligned middle, in fact, some algos have versions for multiples of 8 bytes too just so they don't need to bounds check after each 64-bit item.

Any special reason you can't use SIMD?

State machine is a super general model, pretty much the only thing you're promising is that the subset of state that selects an execution path on the next iteration is bounded. you should generally either choose a more specific model or try to phrase your code more directly. This description sounds like you can just put the filters into a list and fold that list into the list of results (in JS fold is the 2-argument form of Array.reduce)

I'm not convinced that this needs a state machine. State machine is a good model for situations where the way each element is processed and not just the numerical result is dependent on the previous section of the list. Even a typical reduce or fold isn't best modelled as a state machine if none of the information produced by the body selects an execution path in later calls to the body.

We have a similar feature, but implemented more transparently.

There's a search criterion list, and when you hit enter in the search field, it clears the field and adds its content to the criterion list. The contents of that list are introspectable. On any keypress, the search results are produced via all of the criteria, including the contents of the search field, if it's a meaningful filter.