memling

He also wrote a great intro to generating finite automata using Scheme macros.[0] It carries you through step by step and motivates the work well.

[0] http://cs.brown.edu/~sk/Publications/Papers/Published/sk-aut...

> It is a fool's errand to try to evaluate the net societal impact of a particular technology, even with perfect predictive power.

This really isn't true. We regularly do cost-benefit analyses for business; we don't skip them because they're not perfect predictors. We do market analysis and all kinds of customer deep dives to perfect UI/UX and customer response. We've created targeted dopamine-delivery services that are continually refined to maximize impact.

All of this implies a certain kind of ability to evaluate. Looking at tradeoffs and potential uses of technology is well within our abilities, and we should do it. We should be more skeptical of human nature, look harder at the extremes and edge cases, and work towards mitigating the risks. Will it be perfect? No. Will it be helpful? Yes.

> I saddens me to see here these sort of posts.

We should actually be much more proactive about technology: it is not an unalloyed good, and an unwillingness to consider its downsides leads to naive designs. Even leaving bad actors aside (which I don't recommend doing), what works well in a group of 10,000 users may not scale to one billion users. Where one scale has no effect on social cohesion, another scale may have a tremendously deleterious effect.

We've been doing these experiments in search and social for years now. Taking lessons from that and applying it to the next great wave of AI innovations seems like a Good Idea to me.

Saunders collection Tenth of December is great.

Gladwell's 10k hours rule is disputed.[1] Assuming it were true, for sake of argument, it's not 10,000 hours to get good at something. The difference between bad and good is much, much smaller than the difference between being good and great at something. Josh Kaufman argues that 20 hours could be enough.[2]

All things being equal, make a change. I did this more or less accidentally (applied for programming, got turned into a systems engineer) and it was terrific. Feeling stupid is awful, but it's pretty good for you in the end.

[1] https://www.goodlifeproject.com/podcast/anders-ericsson/1

[2] https://piped.kavin.rocks/watch?v=5MgBikgcWnY

> Find something that’s enjoyable to you, and valuable for the world, and put your whole self into it

Sounds like you hit one of two here. Not for the faint of heart. I hope yours recovers soon enough.

> There are much less obscure texts, roughly contemporary with the Bible, which discuss humanistic values such as might be associated with 'Western Civilization'. Some of these have been treated as holy by various religious groups including Early Christians, or as of a quasi-religious importance in the organization of society and its institutions. Many of these were widely read throughout the Christian era and are much easier to see as having influenced 'Enlightenment' and 'Modern Secular' values. For example, Plato's Republic. Any suggestion that the Bible is more important than that or similar works to gain "an understanding of the process of how Western Civilization is formed as it is today" just seems to me to be a hangover from Christian apologetics and bigotry.

Have an upvote!

Can you draw some broader pictures here for the relatively uninformed how you would get a broadly humanistic view from, e.g., The Republic? If that's too broad an ask, could you draw a line from Plato (and others) to the Universal Declaration of Human Rights?

> The Bible is not exactly .. readable?

I can appreciate why people might say this, but modern translations are not that hard, and large parts are narrative and relatively straightforward. Probably a bit of an explainer is helpful, just like any other major work of literature.

As to whether OP would find it worthwhile reading, that obviously depends. But as others point out it is the most important book in shaping Western civilization as we know it...so there is probably some value.

On that subject Luc Ferry's Brief History of Thought has perspectives on several major thought streams. I don't know how accessible it is, but when I read it a few years ago, I found it reasonable, and I don't know much about philosophy.

Educate me, pkease: why no new or delete? Does C++ now discourage heap allocation or is the common practice to use smart pointers, etc.?

> "Lies, damned lies, and statistics"

There's a good book on this called "Damned Lies and Statistics" that I recommend to people.[1,2]

[1]: https://archive.org/details/damnedliesstatis00best

[2]: https://www.ucpress.edu/book/9780520274709/damned-lies-and-s...

> That said, a hydrogen gas station exploded near me, so yeah. There's definitely challenges, and it shouldn't be used where it doesn't make sense.

Why did this happen, out of curiosity? Do you have a linked source?

Can you talk a bit more about the reasons for your deconversion? I'm always fascinated by these stories and it seems like yours includes an interesting elemnt of your parents supplying you with the tools to think critically about what you believe and why.

I more or less diagree with you, but the comment was lovely, and for some reason I find myself craving lobster thermidor.

> robust to noise but not warping in time or frequency domains.

Could you use something like a dynamic time warping algorithm for this? (I'm not super acquainted with the technique and not sure if you could get away with it in the frequency domain used for the matching.)

> C has such a bad wrap with the HN crowd...

why?

> Some U.S. states have a sales tax. Some also have an annual tax on the current assessed value of personal property, like real estate, cars, boats, etc.

My current state has sales taxes (on non-essentials), property taxes that go to the state and the local school district (separate assessments), and an income tax to the state and my local district. The state counts income as earned income and passive income. Federal deductions reduce my tax burden there, but no such deductions obtain at the state or local level.

While it's popular to talk about the tax differential between the EU and US—and there is one—it's probably not quite as big as advertised (at least not in all states).

> Depends. Vigorous debate and experimenting does not come from humility. It comes from conviction. Humility is an end game, you can only reach for it when you accept that you are wrong (at the end of the argument) and likewise, when you win and don’t parade it and acknowledge the merits of the other side.

This isn't true. Humility is accurate self-assessment, and from here can easily spring conviction, experimentation, and vigorous debate. (I know my opinions on subject X are well formed and tested by trial, error, and experience; that is why I hold such and such an opinion and can debate about it with conviction. It's also why I can take corrective measures and understand what others are saying, etc., etc.)

Indeed, humility is basically a prerequisite to adequate debate. One of the problems of society is that we don't get this.

> But how can a humble person accomplish something extraordinary? At some point, you gotta have faith in what you're doing, despite all the opposition.

Humility means accurate self-assessment. People will argue this point, I suppose, conflating humility with meekness (and meekness with weakness), but I think a distinction is both correct and worthwhile.

Therefore, humble people can accomplish extraordinary things as well as (or better than) anyone, perhaps because of their accurate self assessments. This leads to interconnection (I'm good at X, not at Y) and complementarity: I strengthen what is weak in you and vice versa, and we develop in partnership in a way that could not be done separately.

> I take a bus. Canada pays too little for software to own both a car and a house

Can you get comped for working on the bus? Is it even possible? Commutes like this are really killer.

If you work a 230-day year, every ten minutes on your commute is over 38 hours of time. You're spending the equivalent of 17 work weeks in the car, or 37.5% of your 230-day work year.

That's a lot of time.

> I'm using Pinephone as a daily driver

What OS? Does MMS work out of the box for you?

> I love my friends and family and they love me back. However I would avoid conflict at almost all costs in order to please people even if it is too my own detriment.

I've sat on this for a couple days trying to wrap some thoughts up. It's challenging, and I think what I've written below is pretty incomplete. But in any case...

I think these two statements are mutually contradictory, and I think that is an essential insight. (I, too, tend towards peace-making and people-pleasing at the wrong times.)

Love is a commitment that should produce sufficient safety to tell the truth to each other. It's hard to do that, and it's hard to test it. Love is ultimately concerned with the good of the other. This reflects around to you: if you are unable to be truthful with the beloved, you are not best for the beloved.

The well-worn bedrock of civilization--do unto others as you would have them do to you--seems like a good starting point for both loving yourself and your neighbor. It remains fraught with difficulty, though, because we often lack the will to do what is best for ourselves (or even the knowledge of what is best).

I have been learning Racket on and off for the last few years as part of an interest in DSLs. (The documentation is well written, but I confess a certain kind of impatience with it because it is hard for me to get task-specific information when I'm trying to solve a particular problem.)

I recently had to remove parinfer because its inferences in multi-line syntax-parse statements were not great: I had to join a bunch of lines together to put in the parentheses I needed and lost placement. Parinfer's unwillingness to let me balance parentheses manually is a frustration. Am I misusing it?

> Scrivener

Are you aware of any tools that approximate Scrivener under Linux?

To add to those who have said that there are legal implications, I would add anecdotally I was advised by my HR department not to give feedback for exactly that reason. (We had a junior candidate who misrepresented his experience on his resume, and I would've offered some advice about this sort of thing for the future.)

There is also often a very high volume of candidates for particular positions, and it is a challenge to offer feedback, especially when they're not brought in for a live interview.

In my own process, interesting positions result in me searching for hiring managers at the company, sending a cover letter+resume that includes a table of their needs listed in the JD alongside my skills. This tends to get me closer to the interview table and provides an opportunity to exchange enough information for feedback purposes. It's also a bit of a differentiator in terms of resume structure (more than one interviewer has commented on it).

> Who cares about inflation if your rent/mortgage is 3 or 4 times higher?

Oh, for sure: inflation is challenging to measure, and it probably isn't very descriptive for a lot of common situations. (I think the Consumer Price Index is an attempt to help solve this issue. CPI compounded about about 2.1% per year between 2000 and 2021.[1,2] The CPI includes measures of rent, but it also adjusts for seasonal changes.[3])

Inflation and CPI both obscure a lot of ground-level realities for people--the nature of all summary statistics is that they fail to capture the totality of reality in favor of giving us a broad sense for what's going on. Anecdotally my folks moved after their rent increased by 50% when their lease expired. (That was last year.) These are obviously nontrivial increases that have dramatic impacts on people.

[1]: https://www.rateinflation.com/consumer-price-index/usa-histo...

[2]: (270.97/172.2)*(1/22.) ≈ 1.021

[3]: https://www.forbes.com/advisor/investing/cpi-consumer-price-...

> I bet 100k today is roughly 50k in the year 2000.

Mean inflation from 2000-2021 (last year on record) was ~2.24%, including a 4.7% increase in 2021.[1]

$50k inflated at this rate would be approximately $80k.

I'm sure others are more qualified to discuss implications, but consider that while inflation erodes nominal purchasing power, it cannot account for effectively deflationary impacts of technology (for an example, see[2]). Of course, we cannot eat technology, at least not exactly; so we probably want to look at metrics related to how easy it is to live. It might be worth looking at cost per calorie, for example, or for commonly-available proxies (e.g., fast food prices).

[1] https://www.usinflationcalculator.com/inflation/historical-i...

[2] In 1999, hard drives were sized in the hundreds of megabytes and cost more than they do today. See, i.a., https://en.wikipedia.org/wiki/History_of_hard_disk_drives#20....

> [M]y vote behind Guards! Guards! as a great starting point.

That was my introduction, when it came out, but I recommend Hogfather to most as it captures the wide breadth of his writing. There's a lot of great comedy in it, but also some serious and scary moments.

Have an upvote.

> So right off the bat I'm not sure the point you're making.

That 350k yearly salary is a high number. Attainable, sure, but high, and not representative of typical SWE jobs for most people, senior-level included. Not that people cannot or don't make that, but that it's probably not the norm.

> If the upper limit you are willing to pay does not exceed the 90th percentile TC, by definition you are not offering "very aggressive" pay.

What qualifies as "very aggressive"? 99th? How is this determined? What, typically, is a business's approach towards determining salary ranges for positions?

FAANGs perhaps don't have to observe that approach, I suppose, given their resourcing and scale, but in my experience it's done by market comps.

> Second your numbers there aren't conditioning on seniority,

Fair enough: it does have an impact on the percentiles. I can't easily get them from Levels thanks to the relatively coarse granularity, but a couple quick observations:

In the Bay Area, the top 100 salaries for 5-10 YOE ("senior engineer" per Levels) have substantially higher rates of total comp: $710k-2+M (n=1 here). Most of this is not base salary (which topped out short of $400k) but stock and bonus.

Top 100 in NYC are similar, though smaller proportionately (tops out at 1.5M TC, N=1).

Top 100 in NoVA/DC is markedly lower (tops out at $500k TC).

> and additionally given not only population size but proportion of tech jobs available the Bay area and NYC account for the majority of software jobs out there.

With respect to where SW Eng jobs are located, that's a great question. I'm not sure how to quantify this very well, but for grins, I tried searching for senior-level positions on Indeed and LinkedIn.

Indeed found ~33k jobs, of which ~12k are represented in the locations on which I could filter. Of those 12k, 6021 were on the coasts (I counted anything in Washington, California, NoVA, NYC as coastal jobs but excluded lower-cost areas like Atlanta); I realize not all of these are in SV proper, but I think it probably captures the idea of high-cost/high-value markets for SWEs, and it makes the numbers more favorable against my point above. 3087 were listed as remote.

It's not well controlled, but Indeed's numbers suggest about 20% are in SV or SV-lite areas, and about 10% remote.

LinkedIn's job filters are maybe a bit more sophisticated? but don't offer the same sort of numerical granularity. I picked senior+ SW titles in the US (so included Lead, Principal), of which 43k+ hits were returned. The top ten markets they offer for search (Seattle, Austin, Boston, Atlanta, Charlotte, NYC, San Fran, Chicago, Sunnyvale, and LA) account for 11k results, about 25%. Excluding low-cost areas (Charlotte and Atlanta) returns 9K results, also about 20%.

So I don't feel terribly uncomfortable positing that of widely-advertised jobs, the high-value markets might account for 20-25% of the SWE job market. This is not a vast majority. The numbers could be off, of course--these aren't scientific samples, the searches will naturally have an impact on results, etc. But I don't think most people in software are working in SV, and I don't think most seniors routinely sniff a $350k salary.

Doesn't mean that the OP (GP? GGP?) is paying aggressively or that advertising salary or total comp is a bad idea--just that my initial reaction that $350 is a lot--seems like a reasonable take.

> If your upper bound does not greatly exceed that number*, then all of the actually senior folks that I know are not going to be interested in your role. Most of my contacts are outside the Bay.

That's fine--it's still a large number, and most roles simply don't pay it. There are, I'm sure, many openings that can pay sufficiently qualified people quite a lot of money, multiples of $350k. But it's a high number.