melony

50 ms·

601.
by melony 5y ago
It is interesting that recursion is still considered to be so novel on blogs and forums. From what I have seen of current CS curriculums, recursion is usually taught in the first year immediately after introductory CS. It is almost imposs
602.
by melony 5y ago
Might want to explain in the README that Prisma's core is written in Rust. You don't need to install an entire JS stack to use this.
603.
by melony 5y ago
The Go standard library?
604.
by melony 5y ago
The difference is that claiming consciousness may kill the entire field of ML research when non-experts decide to wade in and start lobbying for regulation. You don't want misguided groups like PETA meddling with regulating neural netw
605.
by melony 5y ago
Open source Figma
606.
by melony 5y ago
This is actually a lot more sophisticated than the comments here are making it out to be. With something like this you can basically get software defined cooking. If you combine it with computer vision and a bit of chemistry, you would get
607.
by melony 5y ago
https://probml.github.io/pml-book/ Murphy's should be more than sufficient. Mackay ( Information Theory, Inference, and Learning Algorithms ) and Bishop ( Pattern Recognition and Machine Learning ) are also common
608.
by melony 5y ago
Too much focus on implementation details and not enough emphasis on state of the art theory. A new practitioner should start with probabilistic learning for theory and deep learning on the practical side. A cursory overview of linear regres
609.
by melony 5y ago
You are throwing the past two (three?) decades of research and development in software engineering if you insist on tightly coupling code structure to implementation and rejecting design patterns.
610.
by melony 5y ago
They have no real private space industry worth mentioning, relatively speaking.
611.
by melony 5y ago
OP is saying that the rules engine can't catch a bunch of errors that the ML based NLP can. In other words the ML system has a bigger input domain than the rules engine. I am confused about how they plan to use the rules engine to expl
612.
by melony 5y ago
A correct max heap is very difficult to implement from scratch during an interview. Also, shouldn't it be a min-heap?
613.
by melony 5y ago
But if the NLP is able to flag more errors than the rules engine, how could the rules engine explain the errors?
614.
by melony 5y ago
I never understood the obsession with sending heavily armed police to break down doors for some trivial white collar crime. If there is no immediate threat to life, it can wait. You would think that with all the financial and digital survei
615.
by melony 5y ago
It is interesting that the fundamental parts of grammar checking are still driven by rules engines and GOFAI expert systems. Even the gramamarly core is written in common lisp as a GOFAI app. Most of the modern attention/transformer NL
616.
by melony 5y ago
There actually is proliferation concerns, maybe even worse.
617.
by melony 5y ago
It would be better if e.g. the first 50 can be priced at $5 per user per month and once the growth hits $300, revert to $1 per user.
618.
by melony 5y ago
It's not just Linux that suffers from outdated tray APIs. It is the same on Windows. To create a status tray icon button, you need to call the Win32 Systray API (the oldest Windows subsystem, forget about using WinUI or UWP). The way a
619.
by melony 5y ago
For a copy trading app, what compliance problems do you envision using your service? Is the trader being copied liable if the people copying him lose money (generally speaking)?
620.
by melony 5y ago
Do you plan to offer a more developer friendly pricing?
621.
by melony 5y ago
Reminds me of https://www.businesswire.com/news/home/20190403005265/en/Tra... How does your pricing work?
622.
by melony 5y ago
*time to
623.
by melony 5y ago
I did not have to flesh out my reply, what I meant was in an information theoretic sense. We have methods such as differential cryptanalysis that can operate with very few "hints" to start.
624.
by melony 5y ago
We can break codes by cryptography, a network pre trained on all human languages should be able to derive insights unsupervised.
625.
by melony 5y ago
The issue is not what they can provide, just like the issue is not what Apple brings to the table. The problem is that both are too big and overwhelming dominance poses risks of anti-trust/deplatforming depending on which side you beli
626.
by melony 5y ago
Those payment networks are another relic that needs to be gotten rid of.
627.
by melony 5y ago
Legislate them into allowing sideloading until the day their market share drops below a quarter. Hell, I would be fine with letting Apple automatically wipe phones with sideloaded apps and revoking the user's warranty post legislation
628.
by melony 5y ago
The state (in the general sense, as in the federal government) wants to know where money comes from and where money goes to. Control over capital flow is one of the fundamental levers of modern day political power.
629.
by melony 5y ago
You forgot the most important use case: you want to build fintech services and applications, but you don't want to deal with almost a century worth of compliance and a tech stack built on rotten foundations and middleman rentseeking.
630.
by melony 5y ago
Do bigger cows taste better or are smaller ones the best?

More ›