Working at a still-feels-new-to-me job as Graphics Director for Opinion at The New York Times. Our small team publishes arguments and guest essays supported by visual evidence, like these:
But I'm a believer in asking for help in order to cast a wider net. If you happen to stumble across an obscure-yet-newsworthy dataset, or have a strong feeling about a particular guest essayist that we should be approaching, or can't stop thinking about an argument that's itching you — pitches and tips are always welcome: [my hn username]@nytimes.com
This kind of interactive, bespoke content is one of the main reasons I subscribe to the Times. It feels like NYT truly embraced the concept of being able to tell stories and convey information in a completely novel and exciting format, and I find myself really immersed in these pieces all the damn time. It feels very much like you guys are _leading_ rather than following, which is damn impressive for such an old institution.
Hell, I remember last year's Christmas Cookies piece, which was just this lovely immersive slideshow of tight high resolution videos of different stages of the cookie making process and just thinking "this is just _good_ content."
But this stuff is also powerful because of its ability to inform people in a way visually that more traditional text and even older illustrative modalities would've fallen short with. This deep-dive, guided tour, expando-driven approach is just... awesome. Keep up the great work!
Just in case you happen to know the answer: How does taking tree samples (as in your third link) not harm the tree? It seems inevitable that it would do so, at least intuitively.
My understanding — relayed through one degree of separation from Daniel Griffin, the dendrochronologist who wrote the piece — is that the core samples are very long and thin, and care is taken not to injure the tree and to allow it to heal rapidly.
and how Alan Greenspan went from writing 'Gold and Economic Freedom'[0] to pushing low interest rates that would be rad.
I'm not actually a gold-standard supporter, but given the huge shift what average people can expect economically it would be nice to have this addressed by someone other than the conspiracy crowd.
I work in tech for public media and am it's almost disheartening to know what kind of resources other publications have. We have people spread across everything at once. Do you work for the technology department or for the editorial department?
I've been following the NYTimes tech development from a distance for almost 20 years. A couple of things I've picked up:
1) it helps to be the biggest
2) it helps to have rabid boosters with reputational stakes themselves, like Edward Tufte.
3) the tech group maintains the framework, but each section owns their own section layouts.
4) the tech group made a big splash ... eh, sometime between 2005 and 2009 if I recall, with a very early progressive framework. I wish I could recall the details, but it was spicy when it first landed, like "Wow, that's genius, and it looks so good."
Thank you for making JavaScript a somewhat palatable language. Countless millions of us have to write JavaScript to accomplish the daily drudgery and it's in part to your contributions and impeccable sense of style that it's at least somewhat enjoyable.
People probably think my praise is overblown, but if it wasn't for you I'd probably be using wasm to escape JS and the fight against flash would have been for naught.
I'm sick of having to decide between using cloud software and using local software. Cloud software so often needs subscriptions, and if the company dies I lose access to my data. Local software isn't collaborative. I don't want to email files around to myself, or think about versions.
So I'm building a software platform for local first applications on top of CRDTs. Its called Replica, though we haven't talked much about it yet. I want to be able to:
- Edit any data from one device in my house and have it just show up on any other device
- Share items with other people, and collaboratively edit with them
- Support lots of different applications - including multiple different applications live editing the same data. Like a universal plugin model.
Linux can't compete with cloud software like google docs because anyone running hosted platforms gets punished if the platform is successful. Ideally I'd love to get replica embedded in linux, as an alternative for desktop applications to use to store their state. Then users could open up the same app from different computers and have all their data there, and collaborative editing and things like that would just seamlessly work. I want to be able to open the same file in two different editors and have typing in one show up live in the other as I type.
I want to opensource the whole thing, but we'll probably go with some sort of open core model and charge for our official hosted version (which you want for backup and delivery). I want this project to be financially self sustaining - otherwise I don't think it'll survive. But still opensource enough that people can self host if they want to.
My goal is similar to Joseph's (a platform for local first applications using CRDTs), but the approach is slightly different, as I'm building it based on SQLite synchronization using its session extension (https://www.sqlite.org/sessionintro.html) as the encoding mechanism. I plan to incorporate this sync functionality into my web framework (https://github.com/pkulchenko/fullmoon) to allow any application built with it to become "sync-enabled" with just a couple of additional lines of code.
It's a Collaborative "Second Brain" or Zettelkasten, it's local by default but you can sync your brain with other people. Still in aplha. The local version is free and the collaborative one will cost per sync. So if you only want to use the local version then its ok.
Really cool. I have with git-bug[0] similar properties (offline first, identities, generic "crdt like" base data structure... ). Maybe you'd like to have a look and improve each other's design.
Your idea sounds awesome and I hope I'll hear more about it soon.
> I'm sick of having to decide between using cloud software and using local software
However, I want to point out there's an additional option that is becoming more and more accessible in recent years and that is self-hosting, either on a local server or a rented VPS. Most cloud based applications have several self hosted equivalents at various stages of development.
I really like your idea. Might I suggest monetizing it via a marketplace approach? You completely release your product for free as an Open-source offering. Then you create a hosted marketplace for plugins where you take a commission for each purchase there. (I've written a bit about it on my newsletter https://unzip.dev/0x00d-open-source-business-models/#forecas... - last point)
I really share your thought and sentiment on this matter.
I think an example of successful local-first is email and I'm still using email as my main messaging software. The fact that you can receive and send email seamlessly between different companies and organizations is just amazing. Sadly, however, now most of the people do not have their own domain and their emails are mostly managed by big corporations.
All the best for your Replica application, will keep track on the progress.
I love this approach. I'm trying to bake it in https://acreom.com - a local-first md. editor with sync and real-time collaboration(WIP) + plugin system in future.
This is something I wanted. I found remotestorage.io which seems like a cool idea but I am not sure it has moved much in the last couple of years in terms of adoption.
Unless I’m missing something, Etherpad is just for text editing. Not arbitrary application data like we’re working on. And each etherpad document lives on a centralised server. At least, that was true when I worked with the original etherpad devs in 2011. I can’t find a roadmap for 2023.
Replica will work with native apps first. Data will be local to your devices. Servers are necessary for peer discovery, sync and delivery. But you can still make changes offline and sync later. Or ideally, sync directly between your devices over a local network.
Think about it like GitHub. You can use their website, but the data itself (in git) is cloned onto your local computer - and you can replicate it to other devices however you like. The model for replica will be similar: We’ll run an official website to use for replication, backup and discovery. But your data is, and always will be your own.
I am not sure if this applies, but is the concept similar to blockchain... All the data is with everyone and still within your machine as well. Meanwhile, if the machine is disconnected then reconnect and update with other users.
No; its not like blockchains. The data isn't with everyone. Its just on your devices.
Think about email. Your emails aren't on everyone's devices. They're just on your devices. If you go offline, you can still read and reply to emails. But the emails you send won't actually get sent until you go online again. Its like that; but instead of sending messages back and forth, you can edit things and the changes you make get synced while you're online.
I'm starting a 100% Worker Owned For-Profit residential Trash and Recycling Co-operative
All profits go into the neighborhoods we serve and our bylaws are based on the US Constitution. No investors or different class shares. Management term limits, workers vote for CEO and a cooperative ombudsman (termed) for representation during yearly plan/budget proposal.
The CEO proposes a yearly plan/budget to the collective of ombudsman representatives, which vote to approve or amend. The yearly plan includes all P&L numbers for previous years, open balance sheet, salaries, as well as hiring, growth, acquisition plans for the following year(s)
I'm leaving out a lot of details of course, but the overall gist is that: incentives are aligned across the co-op such that management is incentivized to care for their employees above all, and there is no pressure from investors to grow more quickly than our company can sustain. Basically will not pursue increased margins at any cost.
We will provide the best service at a price that ensures we can fully take care of our employees such that they have a great life and can provide amazing service because they have the time, and trust to care about how we serve each other and our residential communities everyday.
would be interested how this works out. I was working with people this way but it turned out people don't do logical decisions, they decide basic on assumptions which may not be true. somebody with a charismatic personally was always better to sell ideas and decision was not based on evidence.
The same constitution that consistently puts a the party that gets fewer votes into power? If I were to every start a business, that's the last place I'd look for inspiration.
Over the summer I got into genealogy — my partner tasked me with finding her biological parents. I did, and afterward, I took a DNA test of my own only to discover that my mom failed to tell me something (she passed away a few years ago): who my biological father is.
This was a surprise to me, my dad, and many others. Some of her girlfriends knew this situation existed, but never got a name.
And so I spent my summer in what I call "old lady Facebook groups" and learned the tricks and the trades to investigative genetic genealogy. One of the useful tips you can apply to your DNA matches called DNA color clustering — it's incredibly useful: https://www.danaleeds.com/dna-color-clustering-the-leeds-met...
Of course, I didn't want to do this by hand. There were some tools out there that existed but they were "old school" software packages: you know, you install it on your own device! "Worse" yet, I'd have to give it my raw password — no thanks! It started there, and morphed into a lot more, and now it's used by "search genies" and consumers alike: https://sherlockdna.com
There's a lot more that can be done in this realm by both providers and hackers. There's a niche for these "pro tools" that exist but the typical DNA test taker is not after any sort of genealogical exploration that requires substantial effort. There is a very sizable group of people who do, though, and those people, I have found, are very pleasant, kind, and nice to be around. I like surrounding myself with good people, so I'm happy to help them while I scratch my own itch.
All this, and I am yet to find the guy I'm looking for despite having invested over 1,000 hours into finding him. I have, though, found over two dozen "wanted" individuals (not in the criminal sense) and united them with their searcher — at least that part is satisfying.
Have you submitted your DNA results to 23andMe or other services? Is this how you were doing the "color clustering", by matching 2nd and 3rd cousins from matches on those sites?
I have no experience in the matter but presumably there's some research for familial facial recognition? So maybe you could collect family photos or "scrape" facebook local to your area to see if you get any hits? Its pretty speculative but maybe it'd be worth a shot?
I have a little experience in bioinformatics so I'd be happy to help inasmuch as I'm able, if that's something you'd want or need. Regardless I would love to hear more about what methods you've tried (and presumably been unsuccessful in).
Oh yes, I have everything under the sun: Ancestry, FTDNA, MyHeritage, 23andMe. I’ve even had the leading experts in the field helping me and working alongside me on this but it’s really just a case of bad matches.
There’s only so much you can learn from a half-cousin once-removed: her grandfather, born 1911, went to a barn dance, had a good time, and didn’t stay in touch. This man produced my paternal grandma according to my YDNA line — YDNA comes from a male and is passed to males. The YDNA has given me some direction, but no solid answers: I’m trying make a single male from a combination of a male from the most prominent family in 1800s Saugerties, NY, an unknown from Chicago (by way of Bohemia), and two unknowns from two different parts of Grand Forks, North Dakota (both Scandinavians).
And this all happened before the internet of course: 1860-1960 is the range that I am working with.
This h1C1R I speak of is my best match, followed by third-cousins. I’ve exhausted every avenue that’s been visible to me. To the surprise of absolutely nobody, it’s very uncommon to know who casually slept with who generations ago. :)
There’s a lot of census searching (and scraping, and CSVing), newspaper article reading (lots of juicy details in small newspapers) and some luck. I have a tree built out some 1600 people strong. Hopefully with the holiday sales, new matches will come my way that are helpful. :)
Fascinating, I'm sorry your search has yet to yield results but maybe you could try to join up with a company that mainly supports law enforcement as I guess they probably have a robust toolset. I've heard of Othram through acquaintances, may be worth a shot
Since the start of the summer I have been building a SaaS for companies doing inspections and maintenance on separators, a certain kind of device for cleaning water used in gas stations, restaurants, garages etc.
I have been dreaming of living on my own product since I started as a developer 10 years ago and tried a lot of different ideas out but it just wasn't happening.
When I found out that my girlfriend was pregnant with twins in spring 2022 there was this now or never moment. I thought there would never again be enough time available in my life to do something like that.
I decided to quit my job to put a gun against my head. Then I contacted hundreds of people on LinkedIn and other places asking them if they had an idea on a product to build. One guy responded telling me about how far behind in digitalization his industry is and that we could potentially do something here.
I went with it and have been building it since.
As the product starts to mature we have showed it to potential customers and the response has been amazing. We have been promised around $8000 / month from 8-9 companies as soon as we finish a couple of more features and we estimate that there are at least 100 potential customers in our country.
This autumn has been so intense since I'm doing this while doing 40h/week consultation do get money plus having 2 newborn kids at home. I feel very lucky though that I finally got an opportunity to go all in on my own product.
One thing I have learned:
There are so many software developers with both the skills and drive to create a business, but what is super hard is to just come up with an idea. The key is to connect with people outside of our community. The whole world is waiting to be digitalized and our skills are in short supply. The further away from the normal dev community you go, the less crowded will it be and the easier for you to find an opportunity to add value.
Very nice. Having walked down this path: get the signature under the contract or an advance before you build "one more feature".
In my experience "one more feature" is usually a very polite way to say "no", and it can be a sign that you haven't identified the proper solution to their "hair-on-fire" problem yet.
(I've built the exact features prospects requested, and after showing the built feature I usually got another excuse, i.e. "I would also need feature y first", or "not right now", "no budget ATM",... )
If you nail the hair-on-fire problem people should be considering how to finance it or start asking "by when can we use it".
Having been a developer first, I also hugely underestimated the importance of a proper ICP, value proposition etc. Your messaging might be a big part of the reason that people are saying no or yes on the spot.
My old video game from 2004 [1] is now older than I was when I wrote it, and I recently found the C++ source and sprites in one of my archives, so I'm rewriting it in TypeScript as a personal challenge.
The C++ code no longer builds (it's missing proprietary dependencies), and I no longer have the binaries or even the hardware needed to run them (it's a PocketPC game).
If I manage to finish the rewrite, I'll get in touch with the rest of the team, to ask if they'll let me upload it somewhere public.
So far, I'm surprised by how readable my C++ code is. :-)
I've still got my Dell X50V stored away somewhere, if you find the binaries you're invited to contact me. My Gmail username is the same as my HN username.
I'm continuing work on the No Bullshit Guide to Statistics, which is an introductory book on statistics that includes all necessary prerequisites (practical data analysis using Pandas and probability theory math using the computer models from scipy.stats). It's been a many-year project (because STATS IS TOUGH!), but I've had very good progress in the last couple of months of 2022 (finished first few chapters) and now keeping the momentum going into 2023.
It's really hard to explain statistical concepts without bullshit... there are all kinds of formulas and rules of thumb, so it sometimes takes me 2x or 3x rewrites before I find an explanation that is simple enough and free of unnecessary historical baggage. One thing that I have going for me is the ability to use computer simulations to check all the equations, and generate nice visualizations. Seeing the equations + graphs + code in parallel makes stats concepts finally click... at least for the author, I hope for the readers too ;)
If you're interested in checking this out, I maintain a continuously-updated book outline here: https://docs.google.com/document/d/1fwep23-95U-w1QMPU31nOvUn... I also have a mailing list where I send chapter preview PDFs as they become ready (see the end of the gdoc).
This sounds excellent. Will be keeping an eye on this. As I've been spending more time doing data viz recently, I've started to realize how little I remember about stats from college and am having to study back up. Unfortunately, the resources I've found have left much to be desired. My best bet at the moment seems to be going through MOOC courses to essentially retake my stats courses. But your guides here seem much nicer.
> Unfortunately, the resources I've found have left much to be desired.
Yeah most textbooks out there are pretty bad (aim to cover the standard curriculum, which is not very useful for most practical statistical analysis tasks).
Recently however, there has been a growing number of teachers/educators adopting the "modern statistics curriculum" based on resampling methods (bootstrap estimation and permutation tests), so you should look for those if you want a review that is more friendly to learners with a coding background.
I have another google doc where I've saved the best links to learning resources, you might want to check out while waiting for the book (it might take a while!): https://docs.google.com/document/d/1GWyuNC4qCL5ecYSoY6GxvuhO... In particular the talks by Vanderplas and Downey in the section "Tutorials > Resampling methods" are all excellent.
Also, here is a four-part video tutorial + notebooks that I prepared, which gives an overview of all parts of modern curriculum (from data management, through prob theory, to stats modelling): https://nobsstats.com/stats_overview/README.html
My mom departed when I was 19. My oldest son is almost 6. I had a moment last week where I realized that if I applied that scenario, I’m about 1/3 through the time I have with my kids. Obviously it doesn’t work that way, but it was a moment that brought shock then clarity.
Life is fleeting.
So I’m working on “working” as little as I need to.
Good for you! A comment, purely anecdotal obviously, but after seeing my parents for the Holidays I was reminiscing about fond memories and actually (at least with my dad) many of them were around work stuff we did together. He worked on barcode scanners and I thought it was so cool when he would show them to me and explain how they worked. So sometimes, work can be family time too!
Readit News
- https://www.nytimes.com/interactive/2022/10/29/opinion/scien...
- https://www.nytimes.com/interactive/2022/09/08/opinion/urban...
- https://www.nytimes.com/interactive/2022/07/20/opinion/ancie...
But I'm a believer in asking for help in order to cast a wider net. If you happen to stumble across an obscure-yet-newsworthy dataset, or have a strong feeling about a particular guest essayist that we should be approaching, or can't stop thinking about an argument that's itching you — pitches and tips are always welcome: [my hn username]@nytimes.com
Hell, I remember last year's Christmas Cookies piece, which was just this lovely immersive slideshow of tight high resolution videos of different stages of the cookie making process and just thinking "this is just _good_ content."
But this stuff is also powerful because of its ability to inform people in a way visually that more traditional text and even older illustrative modalities would've fallen short with. This deep-dive, guided tour, expando-driven approach is just... awesome. Keep up the great work!
Just in case you happen to know the answer: How does taking tree samples (as in your third link) not harm the tree? It seems inevitable that it would do so, at least intuitively.
My understanding — relayed through one degree of separation from Daniel Griffin, the dendrochronologist who wrote the piece — is that the core samples are very long and thin, and care is taken not to injure the tree and to allow it to heal rapidly.
Here's a write-up from Carleton with more info about how the tree core sampling process works: https://serc.carleton.edu/trex/students/labs/lab2_2.html
Deleted Comment
and how Alan Greenspan went from writing 'Gold and Economic Freedom'[0] to pushing low interest rates that would be rad.
I'm not actually a gold-standard supporter, but given the huge shift what average people can expect economically it would be nice to have this addressed by someone other than the conspiracy crowd.
[0]https://www.google.com/url?sa=t&source=web&rct=j&url=https:/...
1) it helps to be the biggest
2) it helps to have rabid boosters with reputational stakes themselves, like Edward Tufte.
3) the tech group maintains the framework, but each section owns their own section layouts.
4) the tech group made a big splash ... eh, sometime between 2005 and 2009 if I recall, with a very early progressive framework. I wish I could recall the details, but it was spicy when it first landed, like "Wow, that's genius, and it looks so good."
People probably think my praise is overblown, but if it wasn't for you I'd probably be using wasm to escape JS and the fight against flash would have been for naught.
Just a feather in your cap for the new year!
I can see that this isn't your personal design decision, but there is a literal paywall preventing you from communicating with me. This is absurd.
- https://www.nytimes.com/interactive/2022/10/29/opinion/scien...
- https://www.nytimes.com/interactive/2022/09/08/opinion/urban...
- https://www.nytimes.com/interactive/2022/07/20/opinion/ancie...
So I'm building a software platform for local first applications on top of CRDTs. Its called Replica, though we haven't talked much about it yet. I want to be able to:
- Edit any data from one device in my house and have it just show up on any other device
- Share items with other people, and collaboratively edit with them
- Support lots of different applications - including multiple different applications live editing the same data. Like a universal plugin model.
Linux can't compete with cloud software like google docs because anyone running hosted platforms gets punished if the platform is successful. Ideally I'd love to get replica embedded in linux, as an alternative for desktop applications to use to store their state. Then users could open up the same app from different computers and have all their data there, and collaborative editing and things like that would just seamlessly work. I want to be able to open the same file in two different editors and have typing in one show up live in the other as I type.
I want to opensource the whole thing, but we'll probably go with some sort of open core model and charge for our official hosted version (which you want for backup and delivery). I want this project to be financially self sustaining - otherwise I don't think it'll survive. But still opensource enough that people can self host if they want to.
https://github.com/vlcn-io/cr-sqlite/
I think the first product they're building on it is Loop: https://www.zdnet.com/article/microsoft-introduces-loop-a-ne...
It's a Collaborative "Second Brain" or Zettelkasten, it's local by default but you can sync your brain with other people. Still in aplha. The local version is free and the collaborative one will cost per sync. So if you only want to use the local version then its ok.
https://github.com/ilse-langnar/notebook/
[0]: https://github.com/MichaelMure/git-bug
Upspin: https://upspin.io/
Perkeep:https://perkeep.org/
> I'm sick of having to decide between using cloud software and using local software
However, I want to point out there's an additional option that is becoming more and more accessible in recent years and that is self-hosting, either on a local server or a rented VPS. Most cloud based applications have several self hosted equivalents at various stages of development.
Check out this list:
https://github.com/awesome-selfhosted/awesome-selfhosted
Possibly your project falls into the self hosted category in some way?
I really share your thought and sentiment on this matter.
I think an example of successful local-first is email and I'm still using email as my main messaging software. The fact that you can receive and send email seamlessly between different companies and organizations is just amazing. Sadly, however, now most of the people do not have their own domain and their emails are mostly managed by big corporations.
All the best for your Replica application, will keep track on the progress.
Replica will work with native apps first. Data will be local to your devices. Servers are necessary for peer discovery, sync and delivery. But you can still make changes offline and sync later. Or ideally, sync directly between your devices over a local network.
Think about it like GitHub. You can use their website, but the data itself (in git) is cloned onto your local computer - and you can replicate it to other devices however you like. The model for replica will be similar: We’ll run an official website to use for replication, backup and discovery. But your data is, and always will be your own.
Think about email. Your emails aren't on everyone's devices. They're just on your devices. If you go offline, you can still read and reply to emails. But the emails you send won't actually get sent until you go online again. Its like that; but instead of sending messages back and forth, you can edit things and the changes you make get synced while you're online.
All profits go into the neighborhoods we serve and our bylaws are based on the US Constitution. No investors or different class shares. Management term limits, workers vote for CEO and a cooperative ombudsman (termed) for representation during yearly plan/budget proposal.
The CEO proposes a yearly plan/budget to the collective of ombudsman representatives, which vote to approve or amend. The yearly plan includes all P&L numbers for previous years, open balance sheet, salaries, as well as hiring, growth, acquisition plans for the following year(s)
I'm leaving out a lot of details of course, but the overall gist is that: incentives are aligned across the co-op such that management is incentivized to care for their employees above all, and there is no pressure from investors to grow more quickly than our company can sustain. Basically will not pursue increased margins at any cost.
We will provide the best service at a price that ensures we can fully take care of our employees such that they have a great life and can provide amazing service because they have the time, and trust to care about how we serve each other and our residential communities everyday.
What does this actually mean? (Could you link the bylaws?)
I find co-ops fascinating and wish you every success. And that the idea might percolate wider.
The same constitution that consistently puts a the party that gets fewer votes into power? If I were to every start a business, that's the last place I'd look for inspiration.
This was a surprise to me, my dad, and many others. Some of her girlfriends knew this situation existed, but never got a name.
And so I spent my summer in what I call "old lady Facebook groups" and learned the tricks and the trades to investigative genetic genealogy. One of the useful tips you can apply to your DNA matches called DNA color clustering — it's incredibly useful: https://www.danaleeds.com/dna-color-clustering-the-leeds-met...
Of course, I didn't want to do this by hand. There were some tools out there that existed but they were "old school" software packages: you know, you install it on your own device! "Worse" yet, I'd have to give it my raw password — no thanks! It started there, and morphed into a lot more, and now it's used by "search genies" and consumers alike: https://sherlockdna.com
There's a lot more that can be done in this realm by both providers and hackers. There's a niche for these "pro tools" that exist but the typical DNA test taker is not after any sort of genealogical exploration that requires substantial effort. There is a very sizable group of people who do, though, and those people, I have found, are very pleasant, kind, and nice to be around. I like surrounding myself with good people, so I'm happy to help them while I scratch my own itch.
All this, and I am yet to find the guy I'm looking for despite having invested over 1,000 hours into finding him. I have, though, found over two dozen "wanted" individuals (not in the criminal sense) and united them with their searcher — at least that part is satisfying.
I have no experience in the matter but presumably there's some research for familial facial recognition? So maybe you could collect family photos or "scrape" facebook local to your area to see if you get any hits? Its pretty speculative but maybe it'd be worth a shot?
I have a little experience in bioinformatics so I'd be happy to help inasmuch as I'm able, if that's something you'd want or need. Regardless I would love to hear more about what methods you've tried (and presumably been unsuccessful in).
There’s only so much you can learn from a half-cousin once-removed: her grandfather, born 1911, went to a barn dance, had a good time, and didn’t stay in touch. This man produced my paternal grandma according to my YDNA line — YDNA comes from a male and is passed to males. The YDNA has given me some direction, but no solid answers: I’m trying make a single male from a combination of a male from the most prominent family in 1800s Saugerties, NY, an unknown from Chicago (by way of Bohemia), and two unknowns from two different parts of Grand Forks, North Dakota (both Scandinavians).
And this all happened before the internet of course: 1860-1960 is the range that I am working with.
This h1C1R I speak of is my best match, followed by third-cousins. I’ve exhausted every avenue that’s been visible to me. To the surprise of absolutely nobody, it’s very uncommon to know who casually slept with who generations ago. :)
There’s a lot of census searching (and scraping, and CSVing), newspaper article reading (lots of juicy details in small newspapers) and some luck. I have a tree built out some 1600 people strong. Hopefully with the holiday sales, new matches will come my way that are helpful. :)
I have been dreaming of living on my own product since I started as a developer 10 years ago and tried a lot of different ideas out but it just wasn't happening. When I found out that my girlfriend was pregnant with twins in spring 2022 there was this now or never moment. I thought there would never again be enough time available in my life to do something like that. I decided to quit my job to put a gun against my head. Then I contacted hundreds of people on LinkedIn and other places asking them if they had an idea on a product to build. One guy responded telling me about how far behind in digitalization his industry is and that we could potentially do something here. I went with it and have been building it since.
As the product starts to mature we have showed it to potential customers and the response has been amazing. We have been promised around $8000 / month from 8-9 companies as soon as we finish a couple of more features and we estimate that there are at least 100 potential customers in our country.
This autumn has been so intense since I'm doing this while doing 40h/week consultation do get money plus having 2 newborn kids at home. I feel very lucky though that I finally got an opportunity to go all in on my own product.
One thing I have learned: There are so many software developers with both the skills and drive to create a business, but what is super hard is to just come up with an idea. The key is to connect with people outside of our community. The whole world is waiting to be digitalized and our skills are in short supply. The further away from the normal dev community you go, the less crowded will it be and the easier for you to find an opportunity to add value.
In my experience "one more feature" is usually a very polite way to say "no", and it can be a sign that you haven't identified the proper solution to their "hair-on-fire" problem yet.
(I've built the exact features prospects requested, and after showing the built feature I usually got another excuse, i.e. "I would also need feature y first", or "not right now", "no budget ATM",... )
If you nail the hair-on-fire problem people should be considering how to finance it or start asking "by when can we use it".
Having been a developer first, I also hugely underestimated the importance of a proper ICP, value proposition etc. Your messaging might be a big part of the reason that people are saying no or yes on the spot.
May everything work in your favor in 2023!
The C++ code no longer builds (it's missing proprietary dependencies), and I no longer have the binaries or even the hardware needed to run them (it's a PocketPC game).
If I manage to finish the rewrite, I'll get in touch with the rest of the team, to ask if they'll let me upload it somewhere public.
So far, I'm surprised by how readable my C++ code is. :-)
[1]: https://nicollet.net/blog/darklaga.html
It's really hard to explain statistical concepts without bullshit... there are all kinds of formulas and rules of thumb, so it sometimes takes me 2x or 3x rewrites before I find an explanation that is simple enough and free of unnecessary historical baggage. One thing that I have going for me is the ability to use computer simulations to check all the equations, and generate nice visualizations. Seeing the equations + graphs + code in parallel makes stats concepts finally click... at least for the author, I hope for the readers too ;)
If you're interested in checking this out, I maintain a continuously-updated book outline here: https://docs.google.com/document/d/1fwep23-95U-w1QMPU31nOvUn... I also have a mailing list where I send chapter preview PDFs as they become ready (see the end of the gdoc).
Yeah most textbooks out there are pretty bad (aim to cover the standard curriculum, which is not very useful for most practical statistical analysis tasks).
Recently however, there has been a growing number of teachers/educators adopting the "modern statistics curriculum" based on resampling methods (bootstrap estimation and permutation tests), so you should look for those if you want a review that is more friendly to learners with a coding background.
I have another google doc where I've saved the best links to learning resources, you might want to check out while waiting for the book (it might take a while!): https://docs.google.com/document/d/1GWyuNC4qCL5ecYSoY6GxvuhO... In particular the talks by Vanderplas and Downey in the section "Tutorials > Resampling methods" are all excellent.
Also, here is a four-part video tutorial + notebooks that I prepared, which gives an overview of all parts of modern curriculum (from data management, through prob theory, to stats modelling): https://nobsstats.com/stats_overview/README.html
Life is fleeting.
So I’m working on “working” as little as I need to.