HN user

afwaller

130 karma
Posts1
Comments31
View on HN

Anything the cache doesn’t have, causes a cache miss, which passes it through.

So when you request the large file, the CDN (cache) doesn’t have it, so it immediately passes through to the original source. The details of how this is implemented don’t really matter in the sense that they are not going to host the large file(s) in the CDN, so they will always fall through.

Apple is in FAANG. For some reason (likely related to 2015-17 era SWE compensation) FANG just includes Netflix (with amazon facebook and google) and not Apple.

The hacks were not the problem. They engaged in financial shenanigans because they had a bankrupt management culture.

Their decline and fall were the direct result of their human problems. And, briefly, their human problems stemmed from their human problems.

Culture and management matter.

Bitcoin is a pretty reasonable story for how this occurs - first you start on CPU, then GPU, then custom ASICs.

If the financial motivation is there, companies will build custom hardware. Companies like Apple and Google (TPUs) are building custom processors for various purposes.

Sometimes losing that information is very useful however, to arrive at a representation that has a specific meaning.

For example, an x-ray computed tomography (CT) image volume in 3D may be projected into various 2D synthetic planar x-ray projections (digitally reconstructed radiographs).

There are countless situations where projections are very useful.

Apple’s announcing their desktop ARM strategy next week.

Given how many iOS and Mac devices (many Macs have an ARM processor for things like the touchbar) Apple ships, they’re already one of the larger custom high-end ARM licensees. It will be interesting to see how it goes for them.

Lemonade files S1 6 years ago

You only make insurance cheaper by charging risky people more.

Right now it is mostly laws that protect categories of people that keep insurance companies from charging people more.

What’s the plan here, use machine learning in a “hands off” way with a black box algorithm to apply pricing discrimination in a way that a human could not because of regulation?

We had to deal with the sectigo / addtrust expiration. It was annoying because we reached out to our cert vendor and they had claimed it would not affect us (only affecting “legacy” systems or old browsers), but then of course it did affect us. It was an easy resolution, but still annoying.

Of course, we still have people working on our stuff. If it were some abandoned hardware product I could see this being a disaster.

Instead of blurring you should add a significant amount of extraneous information (random noise) and then mosaic (downsample).

If you’d like to have a smooth looking censored image you can then blur the mosaic result to have a smooth transition between the censored and original image.

If you simply blur or simply downsample there’s a significant ability to recover data or iterate over data to recover likely inputs. Other posts have discussed deconvolution, but think of a downsample as a hash - you can build a rainbow table of inputs, easily for numbers, with more difficulty for faces. If you have a limited pool of “suspects” this technique can work well. Just as with hashing, you should add a salt to the image before downsampling or blurring to make recovery of the original input more difficult. In this case the “salt” is random noise.

Zoom probably benefits heavily from access to the lobbyists and lawyers which form oracle’s core product offering.

This "off by one" issue, combined with the problem of words related to latin "mille" for thousand in many languages, is why in money people often use K for units of "1,000", MM for units of "1,000,000", and B or sometimes BB for units of "1,000,000" (BB has no meaning as far as I'm aware, people just copy the style of MM).

Unfortunately it is likely too late to really adopt the metric system for financial transactions. But the K has crept in for thousands.