geekraver

It would be great if Tesla could actually get speed limit detection right for starters, and then slow down the car in time for a new reduced speed limit. Had my MY for just a few days and found that inability pretty annoying. It detected a 40mph road as 60, and a 60mph freeway (which is only ever 60mph) as 70. I’ve had 5 year old Garmin GPS units that do better than that offline. I mentioned this in a local owner group and a bunch others chimed in saying they see this too.

AI is a bogus term. It was well-discredited in the 1980s and has gone through an ill-deserved rehabilitation. Deep neural nets are a step forward for some tasks like vision, but are still just scratching the surface of what really goes on in our brains when sense-making from visual data.

Seems like they’re just trying to confuse things by claiming confusion where there was none.

It’s written by Bob and the others aren’t. That’s a good enough reason for me.

I wonder how many of these verbatim examples are because the training data code itself occurs multiple times in GitHub because it in turn was copied from an upvoted answer in StackOverflow, LOL.

Makes absolutely no sense. The main point of a PIP is to give people time to make a graceful exit; sure, in some cases performance can be turned around but that’s not really what it is about. So you should want the employee to know. Supposedly Bezos is going to be focusing on making Amazon a great place to work; here’s some low-hanging fruit for you Jeff.

And you would have a whole lot of blue marbles.

Good luck using type-based autocomplete to write entire functions for you.

I have reprocessed and regenerated the PDFs. They are mostly more legible now (although a couple went the other way), have color, and are smaller in size. I did need to recreate the repo, which meant a couple of people who starred it would have lost the bookmarks. But the new repo URL is the same. https://github.com/gramster/college-notes

May be of historical interest to someone.

While there may be plenty of people of all races who were not offended by the name, when you operate at the scale of a GitHub there is going to be some percentage who are. Some of them will complain. A company like GitHub then has two main options: change the name, or defend not changing the name. Whichever one they pick is going to cause various forms of backlash from various people, but it’s pretty obvious that changing the name is more defensible and the better long-term approach.

Doesn’t the cumulative error from all of these adjustments just mean garbage comes out the other end? This whole thing seems more like clickbait to give the company publicity.

This is why I require any app to be able to use a hierarchical folder of markdown documents as its store, even if the app is single platform. Currently using Obsidian on the desktop and MWeb on my iPhone, but the docs are just such a collection of markdown so easy to share and no lock-in.

I have pointed out both paid and bait and switch reviews (where the seller uses a cheap product to be able to use a farm for fake “verified purchase” reviews then changes the product). Never been rejected, but maybe that’s because I review product too. Amazon seems to bump up positive reviews though so I suspect not many people see mine unless you’re like me and explicitly focus on reading negative reviews first.

Leave it to munificent to explain things clearly. Every time :-)

You could replace a whole bunch of these with a Synology NAS if you really want to own your own data. DSM has an extensive suite of good apps. https://www.synology.com/en-us/dsm

A magazine called “Reason” publishes a piece which in the second paragraph has a line about some things being “90% bullshit” 🧐

Pylance supports .pyi files.

256MB RAM? How extravagant! My first computer had 3kB.

This is just the nature of “induced demand”. We might expand the power of our computers by several orders of magnitude, but our imaginations don’t keep up, so we find other ways of using all that capacity.

I am old enough to have been writing software since 1976, and still do, although I am mostly a manager of managers now. My view is that dates are unfortunate (but sometimes unavoidable in the real world). But estimates are useful. They help flesh out risk, uncertainty, and ambiguity. They provide a set of approximations against which you can plot a glide path. And of course they must be frequently updated in the face of a changing reality. But that constantly evolving glide path is the “current best known” ship date, and that is very useful information (actually you can go deeper and look at its rate of change to give a better approximation).

It’s the job of engineering management to convey to upper management that these dates are projections only with certain levels of confidence.

“ Note: This book does seem to overgeneralize and oversimplify complex situations, but there are still good points.” describes everything by Malcolm Gladwell.

Lots of good recommendations here; two I would add are "Debugging Teams" and "The Geek Leader's Handbook"

The way this is implemented in VS Code on the desktop may change your mind. See for example https://devblogs.microsoft.com/python/remote-python-developm...

You don't need to import an .ipynb. Just start creating code blocks with #%% and you will get a Run Cell code lens appear.

Whenever you see a bill called “X First” for some X, you know who it’s going to shaft - X.

This is a huge improvement. Supports unittest, pytest and nose.

The irony is he is one of the most closed-minded people around, if you see him on Twitter etc.

You're confusing the Jupyter backend (which manages the Python kernels) with the front-end (which maintains the JSON document that reflects the notebook). You can have different front ends that use different formats than .ipynb.