oakfr

Turns out that it also means fleeing :)

The “cost” of being GDPR compliant is negligible if you simply put static content on your website and don’t track users.

Indeed. Denial of responsibility is a feature of violent communication.

https://blog.chaddickerson.com/2018/10/29/nonviolent-communi...

The 4% rule is wildly optimistic. You're assuming that there exists an investment that:

- is 100% liquid - will return 4% consistently over decades - every single year - post-tax - post-inflation

Good luck finding that one.

So that's why I am saying that the FIRE theory is a theory. A much more realistic number is 2% (and I am still being optimistic here). That means that for someone who needs $5K/month (not insane if you want to have kids), you need $5K x 12 months x 50 = $3M. To get that sum post-tax, you need roughly $5M of income somehow. The set of people who make that amount early enough in their career is very limited.

Thank you for your reply. I found it very profound and insightful.

I wouldn't be surprised if a high number of us is now really doing 3-4 hours of real work a day, filling in the rest with stuff that we'd have been ashamed to do in front of our colleagues (or called for) in the pre-covid world.

This is a good thing. The pre-covid world was mostly a circus in that respect.

As long as the 3-4 hours of daily deep, profitable work are still there, we're good.

Edit: this comment does not apply to small startups where your productivity as an engineer is much higher and it (sadly) makes sense to work for 8+ hours a day.

I've read about the FIRE theory time and again. For some reason, I don't buy it. I fail to believe in the plan that consists in building enough FU money to then "be free".

Free to do what? What was the plan originally?

I think the real meaning of life is to do what you care about today, _with the constraints you have today_.

Experienced manager as well. I highly concur.

Sadly, the "good" solution is much harder to accomplish than one might think.

One way to address it is to take it from the People perspective. Are there people you've really enjoyed working with? Go talk to them, see what they work on, etc.

This may get you out of your current (deep) local minimum. Doing this on your own is really, really hard.

I am with you on most of your points, but definitely not on the "normal and mostly acceptable".

The truth of our domain is that you can get great work done in 3-4 hours a day (I am referring to deep work here, not meetings). In fact, most of us become much less productive beyond that. The remaining hours can/should be spent on useful meetings, reading, learning new things, chatting to people about things, etc.

Coasting from a standup to the next with zero work in between is definitely not normal (_if done on a regular basis_) and the sign that something is not right in your current situation.

You may find it OK right now and for years on end. But you're likely going to pay a hefty bill for this many years down the road. I am not saying you should live Elon's life. But finding something meaningful to do with your life should be a goal, I believe.

> I only invest money I can afford to lose.

You mean "speculate" here, not invest.

People definitely want (and must be able) to invest money that they cannot afford to lose, at least not all in a split second.

Amen.

Indeed. Also, you have plenty of time do to both. Start as an engineer at 22, write code for 15 years. Switch to management at 37. That still leaves you with 15+ years to climb the career ladder.

Using the term "procedural noise processes" instead of "random noise" in the title would have been less click-baity and closer to what the paper is about.

Thank you for your reply!

This article is completely disconnected from what is happening in the machine learning community.

Granted, datasets have grown larger and larger over time. This concentration is actually a healthy sign of a community maturing and converging towards common benchmarks.

These datasets are open to all for research and have fueled considerable progress both in academia and industry.

The authors would be well-advised to look at what is happening in the community. Excerpts from the program of NeurIPS happening literally this week:

Panels: - The Consequences of Massive Scaling in Machine Learning - The Role of Benchmarks in the Scientific Progress of Machine Learning - How Copyright Shapes Your Datasets and What To Do About It - How Should a Machine Learning Researcher Think About AI Ethics?

All run by top-notch people coming from both academia and industry, and from a variety of places in the world.

I am not saying that everything is perfect, but this article paints a much darker picture than needed.

FWIW, this paper is being presented at NeurIPS this week.

This article seems to spawn many threads on online advertising, but the initial comment from the author seems to be more around a system/architecture issue (i.e. why should a browser require backend servers). I am writing this comment here to see what people think about this question specifically.

The thing is: you paid for your bedroom ceiling, your bathroom and your car. You also paid for your computer & screen, but you did not pay for the content you're watching on it.

I am entirely with you on not wishing any ads at all, anywhere. I just think that it's not realistic to remove ads entirely on the open web today.

This should not be an excuse for the way ads are being run today. It is fundamentally broken. Incentives and attributions should be changed.

Same here. My guess is that they constantly AB test/explore captchas to push the envelope of genuine humans users in order to make it as hard as possible for the bots.

The first plot is interesting, but could also be mixing correlation with causation.

I actually find it unlikely that raw financial inequality would be a direct generator of worldwide wars.

Some confounding variables are the root cause, and they are not exposed in this article.

Ong bak was the real movie reference to use for tree climbing.

All that being said, I agree with you: leadership is hard. In 20 years in industry, how many (really) good leaders have I met? Three maybe.

My original point is that the _role_ is a lot more interesting than what the article seems to suggest.

Of course we have all met people who were pompous, arrogant and ignorant.

But there are some good and outstanding leaders out there. If you never had the chance to meet one (or did not recognize them when they were in front of you), I am sorry for you.

It saddens me to see, time and again, the role of leadership painted in such a dire and incorrect way.

Leadership is a lot more than about "shuffling paper" in an "ivory tower".

As a lead (in any decent company), you get to participate in critical design reviews; you get to read and review code; you get to shape the trajectory of others; you get to participate in discussions with the product teams. If you work smart, you get free time that you can use to keep coding. All of this can be extremely rewarding.

You are not serving our domain by painting this kind of picture of leadership. The truth is, you can excel as a software engineer, but you can also excel as a lead. And none is better than the other.

I dedicate this message to the "wet behind the ears" people you are referring to who would never consider the role if they were to stick to this poor vision you're delivering.

Does it apply to jobs running on multiple instances, e.g using dask?

This looks like an obvious approach (in the hindsight of course) to a general problem. I love your pivot.

Congrats on the idea and godspeed. You’ll probably have a lot of interest if you execute well.

What you call the « next word » is actually the next word it has already seen in a similar context in the training set.

The novelty does no go beyond the variability of the training data.

And so deep learning cannot generate futures that are both surprising and general.

In the context of conversations, no doubt that it can generate realistic answers. They will just be regurgitations of the training data. They might seem novel to you because you haven’t experienced all the discussions of the training set. But they won’t be novel to humankind and won’t project it into an actual future.

GPT3 and other deep models cannot predict the future. They can only generate alternative presents.

Incremental ROAS is the right metric in theory, but very hard (or impossible) to measure in practice, in particular for the small businesses with smaller ad traffic. Hence the industry falls back on measurable proxies such as ROAS.

Follow-up question on your comment "my #1 recommendation is to get engineers regularly talking to people who could be your customers".

Do you maintain this direct connection between customers and engineers today, and if so, how?

How does the interplay with product/growth teams work?