I’m not a fan of writing front end code with Python or any language other than JS/TS. How deep can I customize or build feature rich UI? I still think the best combo is Rails/ Django API with Nextjs client
With PySaaS, your app's frontend is compiled down to a NextJS app, and FastAPI is built-in for the backend.
UI components are fully-customizable, and are actually wrappers around React components using Chakra UI. If you want to get fancy, you can even wrap your own components in three simple steps.
Neither Finnish or English is my mother tongue. I spent 8 years in Finland and the rest in the US as an adult. It’s really hard to learn both languages at the same time because of little similarity. My English suffers while I can start conversation in Finnish. Now English is my main language and I couldn’t remember much Finnish.
You didn’t answer the parent post. Your point was, in game theory, there’s no benefit in attacking the network or the loss is huge that it doesn’t worth the the attack. The parent post gave the counter point that there could be a benefit which we haven’t thought. If Dr Evil is heavily invested in 2 network, he might destroy one to focus on the remaining. The chance of the attack is low but it is not zero
I addressed it in the first paragraph of my comment above. Let me quote from it here: The theoretical attack argument is indeed "applicable, not just to Algorand and other block chains, but also, more generally, to ANY transaction platform" -- VISA, Mastercard, Amex, ACH, Fedwire, etc. No one disagrees that it is applicable, i.e., a theoretical threat, to all transaction platforms.
Now, if you think such an attack is an important problem for block chains, then you must also think it is an important problem for all legacy transaction networks. Yet we're all comfortable using our credit cards and bank accounts every day, and for virtually all practical purposes, we don't worry about a "Doctor Evil scenario." Why should we think and behave differently for block chain networks?
Moreover, as I wrote before, in practice, legacy transaction networks (like, say, regional VISA networks run by 100-year-old banks) are easier and cheaper to attack. If the Doctor Evil scenario were a real threat, it would be more profitable for him to target one of the legacy networks!
I like this point, but (as I think you are saying as well) it’s true for POW as well, right?
Fundamentally, if there are off-chain incentives to destroy the value of a given blockchain, much of our reasoning about the game theory doesn’t hold up.
Sorry, it's like a way you can write browser extensions, but not have to worry about the limitations of the various browser vendors or their distribution channels, while delivering the "extended browser" by embedding it as a plug-in a web app. In essence, it's just a browser-as-API that you can embed and run inside a regular web app. This is possible because your web app only hosts the client-side interface which connects to a back-end of a headless browser running in the cloud or your server somewhere.
The no-frills, "just the browser" product, is normally called "remote browser isolation"[0] (or, RBI), but ViewFinder lets you build custom browsing extensions on top of this. So you can embed such a remote browser in your web app to encapsulate some off-site 3rd-party functionality that you nevertheless want to include in whatever process is part of that user story or task, for example, submitting a document that you just generated with the embedding web app, useful in the case where the 3rd-party provides no API to do so, making it more fluent than giving someone a list of instructions to then go off-site to follow, or doing it yourself "manually".
You can also use this interface to plug in to some existing headless browser, such as one running an automation job you have set up with puppeteer for observability and intervention if this bot gets stuck--for example, on a CAPTCHA.
One of the browsing extensions I built atop the basic RBI is "co-browsing" so you can have multiple people connect to the same cloud-based browser, from anywhere around the world, and view and interact with it simultaneously. Maybe two friends can watch a podcast at the same time and chat with the built-in chat? But mostly people are interested in this kind of thing for delivering live, real-time customer support to people who are using their app, while they interact with the client over the chat or a telephone. Ditto for training.
From a more meta level, it's a way to craft any programmable experience atop a browser, with somewhat more power that browser extensions (tho offset by the fact that my API is not well documented and can be improved for this purpose).
You run a browser in computer A, but you use the browser on computer B. How that works is B sends your inputs to A, and A sends the pixels back to B for display. B is in effect a dumb thin client.
I need to hear the story from the other side. But if this is true, this is a horrible boss to work with. I’m surprised how OP could endure working for more than a year. There’s no bad job, just bad manager. It’s quite easy to look up this person in Linkedin. Hint: his initials are reversed
You’re describing credit card fraud and more often the banks eat up the loss and issue your credit back. They don’t bother investigating further then write off the loss in tax. Wire transfers are usually available to withdraw immediately so it’s really hard to recover. Some wire frauds take months to recover via court system. That’s why real estate wire fraud is popular and more complicated over time. Overall crypto is not as bad as you think.
In Europe people between themselves and businesses use what you call wire in America. The credit card charge back is very much American concept, I guess the fees pay for that insurance.
Where in Europe do you live? Credit card chargebacks are an extremely well-known and common concept in the UK, and I've hardly ever used personal wire transfer to pay businesses except for auto-recurring transactions like rent, insurance, utility bills.
They don't, they pass it on to the merchant. This is why working with credit cards or paypal is a pain for a lot of merchants and that's why you as a user have to sometimes jump through hoops to just verify yourself. Merchants have to do this because they know that they will be paying in case of you getting defrauded.
Banks rarely eat credit fraud; they more often than not make the merchant eat the fraud instead, in my experience.
The merchant usually CAN make the bank eat the losses, nit that requires pre-approving each transaction and higher fees, so many merchants skip it (at least until bitten hard)
It could be huge or tiny depending on the size of the mean and standard deviation of the effect you're looking for in comparison to the mean and sdev of unrelated variations.
Given that the QOI is a single multinomial proportion, an unweighted[1] random sample of the population has a maximum[2] ~3.1% margin of error on either side.
You could argue the QOI is something else (the sum or difference, or even the ratio estimator reported in the title). Margins of error on ratio estimators are quite a bit larger because of numerical properties of ratios.
Since there is interest, I ran a quick simulation[3] to derive the CI of the ratio estimate above ("5x"). By Monte Carlo and using standard assumptions, the confidence interval on the "5X" quantity of interest is 4.51 - 6.82x calculated via the quantile method.[4]
The confidence interval is not appreciably smaller in magnitude if you double the sample size, we're well past the point of diminishing returns.
[1] Weighting would induce a design effect, but here I would suggest that the real problem is a poorly defined population and sample frame, not sample error. This is the real issue here: we have zero reason to believe these 1,000 people are a random sample of any interesting population.
[2] For a parameter \theta = 0.5 -- uncertainly decreases as the statistic becomes further from 0.5. But classical MOEs overreport like this.
[4] Similar results come from taking the numerical sd of the sampling distribution and calculating a NACI about the mean, I just did the quantile trick because it's less code.
