HN user

antome

465 karma
Posts5
Comments61
View on HN

As someone who has used both PyTorch and TensorFlow for a couple years now, I can can attest to the faster research iteration times for PyTorch. TensorFlow has always felt like it was designed for some mythical researcher that could come up with a complete architecture ahead of time, based on off-the-shelf parts.

This doesn't have much to do with the algorithms, and is more to do with the engineering decisions that went into AlphaGo and AlphaZero. They are designed to play one combinatorial game really well. With a bit of additional efffort and a lot of additional compute, you could expand the model to account for multiple rule / scale variations, maybe even different combinatorial games.

Key part of the article: " According to these upcoming updates, npm will ban:

Packages that display ads at runtime, on installation, or at other stages of the software development lifecycle, such as via npm scripts.

Packages with code that can be used to display ads are fine. Packages that themselves display ads are not.

Packages that themselves function primarily as ads, with only placeholder or negligible code, data, and other technical content. "

I wonder where they will draw the line with the last point.

One thing Microsoft has really nailed, is the location diversity of Azure datacenters. They have multiple locations in South Africa and UAE, while also having regions in multiple cities across Australia, Japan and India. If I wanted to launch a worldwide product, why would I want to go with AWS, Google or IBM?

It's just a matter of following the money. A bigger and bigger slice of Microsoft's income is coming from Azure, and an increasing proportion of Azure users are running Linux. This gives them strong incentive to be the business leaders in open source software.

The other increasing slice is services, where they get you to buy stuff from their storefront, or subscribe to their game pass, or office 365, or make personal interactions with Bing and Windows (so they can sell targeted ads). This naturally gives Microsoft the incentive to know everything about you.

To draw some analogies, Japan has some very conservative people in power, but that's more to do with the election systems that give more power to monolithic parties that attract rural voters. When you look at the proportional vote (Japan has FPTP + a non-reproportioning proportional segment), Japanese people are on average much less conservative than their elected government would make it seem.

Twitter seems to be in this weird space where they have enough usability positives over other social media (pretty broadly accessible, doesn't pester you to log in, you can post NSFW content without it getting taken down) that I guess people just force themselves to use it, even if a blog post would be better formatted.

Affordable RISC-V systems will only happen if they reach a comparable level of mass production as proprietary systems. What you will more likely see for now is a gradual increase in the number of components on GPUs, HDDs etc which use RISC-V instead of ARM or something else. The ISA is only a part of the cost equation.

Similar to Shiny is Bokeh for python, which appears to be trying to replicate the Shiny experience. While you can only get so much control without resorting to JS injections, it's pretty good for [select options -> search -> visualize result] type of problems.

The first thing I do when I get a new Android phone, or update the OS, is disable all animations. As a consequence, my Nexus 5X is in some respects more responsive than many flagship phones today. Sometimes animation can feel nice, but after using a piece of software more than a few times, I'm much more interested in responsiveness.

What's even better (for AMD) is that they can now design a single processor, and scale it across their entire lineup from R3(?) to EPYC. This suggests that they could incrementally improve their entire lineup annually, or deliver huge boosts every couple years.

The are two arguments I see: the first is that if the minimum wage is higher, the cost of living increases, leading to the same problem as before. Of course this assumes the two are perfectly correlated, which is a flawed assumption.

The second is that if local humans are too expensive, people will use local robots or outsourced humans instead. While technically true, this is inevitable. Lowering wages might slow down the incoming problem, but it will always eventually exist regardless of local wages.