enjoylife

It’s essentially a question of did you take a look the code? There is comments in the codebase directly where the config values in question are being used. Eg.

> # repetition penalty from CTRL paper (https://arxiv.org/abs/1909.05858)

Now I’m assuming a base rate of knowledge for this to help, but in general I find diving into the code paths of open source models, usually is a good use of time.

> A human made the decision to widen the guardrails. There aren't a ton of people allowed to do that…

Assuming this is the case, where folks can override during an incident like this, it sounds like the simulation tooling doesn’t have a human loop interface to demonstrate potential effects post override.

Does that type of simulation interface exist in the industry? And would it have helped?

Probably because the raw data is often hidden from the readers so it’s hard to corroborate a stories statistical narrative.

Here is the data which backs up the majority of these low effort cancellation related news articles.

https://public.tableau.com/app/profile/flightaware/viz/Airli...

This is the mental leap folks need to be making, the question is when will this type of configurably factual large language model be research benchmarked and productionized.

I have read many similar comments to the grandparent, all saying the same thing, “I tried it, and it turned up wrong facts” all the while they forgot that they were deeply engaged with the content it was producing. Just think of how much our attention is monetized already, money will be poring into this space from here on. It’s only a matter of time.

Not sure if "wildly emotional" is fair to them either.

> In fairness to Meta: the FTC is the one who initiated this fight, leaving them with the burden of demonstrating it isn't behaving "anti-competitively".

But I agree, the post does seem similar to an individual trying to get attention for their cause, ie. 'Google locked me out...'. To me the tone is probably trying to help sell their narrative of them being this small thing not worth subpoenaing.

It’s 900 pages long, I don’t have a good point of reference so I’ll ask. Is this a small or large reference doc compared to alternatives?

Or phrased another way. Will no name, cheap devices ever properly implement this spec?

> Overall employment of software developers, quality assurance analysts, and testers is projected to grow 25 percent from 2021 to 2031, much faster than the average for all occupations.

10 years is a long enough time horizon in tech for any number of inflection points to occur. My bet is this swings heavily in the downward direction as more specialization in the industry occurs, the usage of the title "software developer" drops, but the overall usage of software and "developers" increases e.g. small businesses using low code tools

Its pattern matching to the extreme with varying amounts of; sequential changes of pictograms and constraints across the rows or columns. What a joke.

A better version of this for our HN audience would be to measure the Lines of Code for our programs to reproduce such a test.

They are betting that hardware combination is the fastest to market, given the constraints of today software.

When the ml stack is capable of leveraging purely camera sensors, Cruise and others like them, own the fleet and can swap out the hardware. Tesla does not "own" the fleet per se. So perhaps its different bets on which cars will still be on the road when the ML threshold is crossed.

I don't know what applications they are developing but an application only needing two custom two hooks, would be a quite trivial one. Not to mention the hook in the article extendState is basically bringing back the Component `this.setState` semantics but not much else.

In a car centric environment such as the US, there is a large burden society pays for when it comes to under age drunk driving. The harm from a car related fatality is felt immediately. It’s hard to compare that to the delayed and difficult to measure societal impact when kids play too much video games.

Haven't seen this automated, but I have seen internal issues raised by folks within large orgs when their extension started breaking things. But it's only once you have 50+ to 100's of engineers working on a product, each with a subset of extensions installed, that you can rely on the cross product of engineers and their installed extensions for realistic coverage.

A cursory glance shows the output of this project is simply bounding boxes for multiple hands, while the handpose project you mentioned is tuned for a single hand but it provides the key points for fingers and joints. I sense that handpose is probably far more useful for anyone going beyond a prototype.

I wonder how Tesla will incentivize the employees who joined circa 2016~2017 to stick around. As they would be fully vested with offers given before the 700% increases. They probably have base salaries in the 100k range but would have equity in the 10m. I could foresee some pretty big turnover as those employees practically retire or head into angel investing. I mean it’s not like Tesla’s Hr is going to refresh their equity anywhere near what they already have.

It's too soon to have a deep opinion on the rfc, but just looking at the examples and thinking ahead...

if folks start using this approach of leveraging a db client in their components, I hope they are embracing dependency injection (DI) somehow, otherwise their codebases are going to become stupidly complex to test.

For an example of how (React + Server Side only plugins + DI) works today, take a look at how its implemented within fusionjs https://fusionjs.com/docs/references/creating-a-plugin/depen...

I agree and not saying they would get to it, but considering that was pre-COVID, I would imagine/hope they might reprioritize.

If your commute is short, say sub 10-15min, enjoying such amenities would not be mutually exclusive to a good work life balance.

I marvel at the jump in BLEU or other measures but I'll second the sentiment, it alone is not showing we are making leaps toward what we need. Yes its a large gradient step minimizing our error, but is it really in the right direction? However I will admit GPT-3 being directed by some yet to be invented causal or counterfactual inference model might be the something which defies my expectations.

Ditto on the not a lawyer aspect, but this seems pretty straightforward:

   You may make use of the Licensed Work, provided that
                      you may not use the Licensed Work for a Document
                      Service.

                      A “Document Service” is a commercial offering that
                      allows third parties (other than your employees and
                      contractors) to access the functionality of the
                      Licensed Work by creating teams and documents
                      controlled by such third parties.

I use Go daily and one of the aspects which is always a challenge is equality checking. From the post,

> There is one subtle way a new field can break user code unexpectedly. If all the field types in a struct are comparable—meaning values of those types can be compared with == and != and used as a map key—then the overall struct type is comparable too. In this case, adding a new field of uncomparable type will make the overall struct type non-comparable, breaking any code that compares values of that struct type.

From experience, this is non-trivial to avoid in a long lived codebase.

It’s telling that 99% of the codegen tools used in Go, e.g. protobuf, thrift, Gorm, etc. all output an isEqual method of some kind.

Your right, they removed that option, https://formulae.brew.sh/formula/ffmpeg. I guess ffplay is built with it by default now.

Sure, not sure what gstreamer plugin/sink would create a loopback device, but this plays as well.

   gphoto2 --stdout --capture-movie | ffmpeg -i - -vcodec rawvideo -pix_fmt yuv420p -threads 0  -f matroska - | gst-launch-1.0 fdsrc fd=0 ! decodebin !  videoconvert ! videoscale ! autovideosink

The USB mode has to be changed. Set it to “PC Remote” mode. https://helpguide.sony.net/gbmig/44840601/v1/eng/contents/TP...

It works for Mac

  brew install gphoto2
  brew install ffmpeg --with-ffplay
  gphoto2 --abilities
  # Abilities for camera             : Sony Alpha-A6300 (Control)
  # ...
  gphoto2 --stdout --capture-movie | ffmpeg -i - -vcodec rawvideo -pix_fmt yuv420p -threads 0  -f matroska - | ffplay -

I'm piping it to ffplay,so this will at least let you test your camera or you could also use it in OBS as a window source. Also, make sure your cameras usb mode is not set to "mass storage" but to a “Remote Camera Control”.

Building a real-time vídeo pipeline for recording and augmenting paintball matches from multiple viewpoints. Basically transforming and combining the scenes into an overhead view with markers for remaining players. It’s a good break from the day job and it has me diving deep into the math behind computer vision.

Upon completion of the interview the problem would indeed look like the result of the leetcode problem. However the take away from the op is in how the question has a narrative. It enables a dialogue over time. Each sub problem provides n ways of exploring a solution. Each problem providing a sounding board for the candidate to pronounce their thoughts. Such questions are effective at providing more signal(read as talking out loud) from candidates.

One big difference is in sports there is a common higher order set of rules which are already established and everyone follows. Contrast that with the outside world where many higher beliefs, e.g. religion are no where agreed upon.

Fantastic paper on containerization of crops. I had no idea this type of analysis even existed

I can imagine the potential advisors/pr folks, having helped shape that type of image. It’s hard to be in the public spotlight and not have PR consultants spamming you.

Considering the space race was abstractly, a push for strategic dominance of a physical domain newly enabled by technologies. I think it’s more likely this time around we will see a push into newly opened digital domains, e.g ml based cyber warfare.