"This necessarily means relinquishing some control over our models. But this compromise is inevitable. We cannot hope to both drive innovation and control it." (2023)
In light of history, they went with the third option.
I think Google wasn't the only one to recognize that availability of inexpensive commodity components was a threat to AGI services. Early on, we also saw a push to release "AI Ready" devices, which have all been shelved, along with a rebranding that calls minor appliance-like systems with some native capability "AI supercomputers".
We were poised for local AI for so many years, and then suddenly... we weren't.
I'm not claiming a handful of companies colluded to manipulate the availability and pricing of commodity components to stifle competition with local/on-prem usage and force people onto services they controlled, but if that's what they intended to do, what might they have done differently?
i like building cool things so i built read rare - the rarest docs and books you've probably never read.
some of the best writing about building things was never meant to be public - steve jobs' agenda email for apple's secret top 100 offsite. facebook's little red book. the 1968 memo intel used to raise its first money. sequoia's "R.I.P. good times" deck.
most of it is buried in court exhibits, dead links, and archive scans, so i put it all on one shelf.
what's the rarest doc or book you've ever read? drop it in the comments and i'll add the best ones to the shelf.
I don't think 'rare' is the right word here. Rare means you can't find it easily, not 'what you have missed or don't know about'. But then what is the word for that? English doesn't have it, afaik. Perhaps 'unexplored treasures'?
Like so: "Less-known treasures you probably haven't read" but this doesn't have a zing to it.
Or? "A collection of useful tech books you've probably missed" sounds more accurate, though not sensational.
'Rarely read' might be a better expression rather than rare. At least compared to 'Cracking the Coding Interview', 'Structure and Interpretation of Computer Programs' etc.
What’s meant by “rare” here? It seems like there might be a language barrier thing going on. There’s nothing rare about, say, John Sculley’s autobiography; I can buy it on Amazon new for about $20 or used for about $5.
The one on Word was very, very interesting. Seems like there were a lot of jerks running the Opus project! I checked them out, and at least on of them died in a plane crash, and now has a foundation.
my goal was to add stuff that is hard to find or that most people have never seen or heard of.
i went through a site that aggregated 40k+ book recommendations from hacker news, and the books i added weren't listed there. so i assumed most people hadn't read them and added them.
This is great, I really enjoyed it. Thanks! The letters and memos are especially interesting. It’s fascinating to see what people were thinking before they knew how things would turn out.
Akio Morita’s book isn’t rare. It was hugely popular in the 80s and 90s and saw another spike in interest after the release of Chip Wars (highly recommended).
Interesting, but not what I was expecting. I have a small book collection, mostly from used book stores around Boulder, CO. I think the rarest and most interesting ones I have are:
"Music By Computers", published in 1969. Before books had websites, CDs, or floppy disks, they had "flexi discs". Covers composing and generating music on computers in the 60s. [0]
"Parallel Logic Programming" published in 1991. Covers implementing Prolog and other logic programming languages on multi-processor systems in the late 80s and early 90s. [1]
"Strand: New Concepts in Parallel Programming", published in 1990. Documents a language that's more-or-less a stepping stone between Prolog and Erlang. [2]
"LISP LORE: A Guide To Programming The Lisp Machine", published in 1987 and is an intro to programming on Lisp Machines. I found a mint condition copy at Goodwill for $1. [3]
"Elementary BASIC: Learning To Program Your Computer In Basic With Sherlock Holmes" published in 1982. Write BASIC code to help Sherlock and Watson solve cases. They just don't make computer books like this any more. [4]
"Man And The Computer" published in 1972, written by John Kemeny, one of the creators of Basic. Predicts the rise of home computers, the internet, Wikipedia, online news, and a bunch of other modern technologies decades before they actually existed. [5]
okay these are actually rare haha. the sherlock holmes BASIC book sounds fascinating. thanks for sharing! will check these out and add them to the shelf.
quick update: added all the books you recommended except music by computers. it's on internet archive, but only people with a print disability can access it.
The Sherlock book was written by Henry Ledgard, a legendary writer and programmer. There's also a Pascal version, which is better because Pascal is a real programming, more capable than Basic-- and doesn't cause you to unlearn bad habits.
Sorry I only recall that someone shared this book on X and I said it’s an interesting book and I love it, and then Sinofsky said the book is inaccurate about early MSFT.
Once I was interested in writing SCP(not sftp) on my own. Good luck finding the spec. I happened upon a random Oracle doc that got me 90% of the way there, and even that felt like a rare treasure.
I'm glad you're excited about what you "built" but it's wildly overproduced as is, and inconsistent in how the animations work. No rhyme or reason to what you've decided to put on the site either, other than "these are some big names that hopefully people will recognize". It reeks of slop.
I’m sure a throwaway account created at the same time as this comment/post and has only commented on this post really cares about the site guidelines and wasn’t created to boost the content.
fair, i probably got carried away with the animations. will fix that.
as for the books, i was just trying to collect stuff that's hard to find or that most people haven't come across/read. but i can see why the selection feels random. will work on that too.
https://newsletter.semianalysis.com/p/google-we-have-no-moat...
"This necessarily means relinquishing some control over our models. But this compromise is inevitable. We cannot hope to both drive innovation and control it." (2023)
In light of history, they went with the third option.
I think Google wasn't the only one to recognize that availability of inexpensive commodity components was a threat to AGI services. Early on, we also saw a push to release "AI Ready" devices, which have all been shelved, along with a rebranding that calls minor appliance-like systems with some native capability "AI supercomputers".
We were poised for local AI for so many years, and then suddenly... we weren't.
I'm not claiming a handful of companies colluded to manipulate the availability and pricing of commodity components to stifle competition with local/on-prem usage and force people onto services they controlled, but if that's what they intended to do, what might they have done differently?
Certainly not rare, but consistent with the readings on the shelves: "Thoughts on Flash" (https://web.archive.org/web/20170615060422/https://www.apple...)
this def belongs on the shelf. added!
hi friends!
i like building cool things so i built read rare - the rarest docs and books you've probably never read.
some of the best writing about building things was never meant to be public - steve jobs' agenda email for apple's secret top 100 offsite. facebook's little red book. the 1968 memo intel used to raise its first money. sequoia's "R.I.P. good times" deck.
most of it is buried in court exhibits, dead links, and archive scans, so i put it all on one shelf.
what's the rarest doc or book you've ever read? drop it in the comments and i'll add the best ones to the shelf.
Which one do you think are the most useful to read?
Not the author, but I thought the Word Postmortem was fascinating.
depends on what you're interested in. i'd start with intel's original funding memo, facebook's little red book, and the word postmortem.
Here's one for you to consider: https://www.wired.com/2000/04/joy-2/
I don't think 'rare' is the right word here. Rare means you can't find it easily, not 'what you have missed or don't know about'. But then what is the word for that? English doesn't have it, afaik. Perhaps 'unexplored treasures'?
Like so: "Less-known treasures you probably haven't read" but this doesn't have a zing to it.
Or? "A collection of useful tech books you've probably missed" sounds more accurate, though not sensational.
yeah but "unexplored treasures" sounds a bit weird for some reason
i think the best description is "a collection of interesting tech books and docs that most people haven't read or even heard of"
If they were truly rare and/or valuable, then I would be honor-bound to submit them to OpenAI for immediate assimilation and destruction
Social media has a neologism for ya: ICYMI
> English doesn't have it, afaik. Perhaps 'unexplored treasures'?
hidden gems?
'Rarely read' might be a better expression rather than rare. At least compared to 'Cracking the Coding Interview', 'Structure and Interpretation of Computer Programs' etc.
What’s meant by “rare” here? It seems like there might be a language barrier thing going on. There’s nothing rare about, say, John Sculley’s autobiography; I can buy it on Amazon new for about $20 or used for about $5.
Yeah, my thoughts too. Though, I will admit: I have never read a number of these. So, perhaps, the author is more perspicacious than we think.
that's what i was going for. some are genuinely hard to find, others are just great reads that most people haven't come across/read
The one on Word was very, very interesting. Seems like there were a lot of jerks running the Opus project! I checked them out, and at least on of them died in a plane crash, and now has a foundation.
my goal was to add stuff that is hard to find or that most people have never seen or heard of.
i went through a site that aggregated 40k+ book recommendations from hacker news, and the books i added weren't listed there. so i assumed most people hadn't read them and added them.
https://hacker-recommended-books.vercel.app/category/0/all-t...
also happy to remove anything that doesn't serve the purpose of the site.
undercooked
I wouldn’t call it “tech”, nor rare. But maybe that’s why I am not in marketing.
This is great, I really enjoyed it. Thanks! The letters and memos are especially interesting. It’s fascinating to see what people were thinking before they knew how things would turn out.
I yearn for internal memos because they reveal the first-person craft, my reading list just got reshuffled!
Akio Morita’s book isn’t rare. It was hugely popular in the 80s and 90s and saw another spike in interest after the release of Chip Wars (highly recommended).
I get the sense that the word rare shouldn't be used here.
Rather this is a nicely presented collection of documents, emails and books that the author thinks are worth looking into.
Maybe undervalued would work better?
apologies should've worded it better. i think the best description is
"a collection of interesting tech books and docs that most people haven't read or even heard of"
or
"a collection of tech books and docs that i find interesting"
Interesting, but not what I was expecting. I have a small book collection, mostly from used book stores around Boulder, CO. I think the rarest and most interesting ones I have are:
"Music By Computers", published in 1969. Before books had websites, CDs, or floppy disks, they had "flexi discs". Covers composing and generating music on computers in the 60s. [0]
"Parallel Logic Programming" published in 1991. Covers implementing Prolog and other logic programming languages on multi-processor systems in the late 80s and early 90s. [1]
"Strand: New Concepts in Parallel Programming", published in 1990. Documents a language that's more-or-less a stepping stone between Prolog and Erlang. [2]
"LISP LORE: A Guide To Programming The Lisp Machine", published in 1987 and is an intro to programming on Lisp Machines. I found a mint condition copy at Goodwill for $1. [3]
"Elementary BASIC: Learning To Program Your Computer In Basic With Sherlock Holmes" published in 1982. Write BASIC code to help Sherlock and Watson solve cases. They just don't make computer books like this any more. [4]
"Man And The Computer" published in 1972, written by John Kemeny, one of the creators of Basic. Predicts the rise of home computers, the internet, Wikipedia, online news, and a bunch of other modern technologies decades before they actually existed. [5]
[0] https://www.amazon.com/Music-Computers-Heinz-von-Foerster/dp...
[1] https://mitpress.mit.edu/9780262200875/parallel-logic-progra...
[2] https://archive.org/details/strandnewconcept0000fost
[3] https://www.amazon.com/Lisp-Lore-Guide-Programming-Machine/d...
[4] https://www.amazon.com/Elementary-Chronicled-Learning-Comput...
[5] https://archive.org/details/mancomputerbyjoh0000john
okay these are actually rare haha. the sherlock holmes BASIC book sounds fascinating. thanks for sharing! will check these out and add them to the shelf.
The Sherlock Holmes book is definitely one of my favorites!
It looks like it's on archive.org, but needs an account to see the whole thing: https://archive.org/details/elementarybasica00ledg/page/n291...
quick update: added all the books you recommended except music by computers. it's on internet archive, but only people with a print disability can access it.
thanks again for sharing these!
The Sherlock book was written by Henry Ledgard, a legendary writer and programmer. There's also a Pascal version, which is better because Pascal is a real programming, more capable than Basic-- and doesn't cause you to unlearn bad habits.
4 is also available on the archive! https://openlibrary.org/books/OL4272985M/Elementary_Basic
[0] also scanned in here: https://archive.org/details/isbn_471910309/mode/2up
[1] here: https://archive.org/details/parallellogicpro00tick
[3] https://archive.org/details/lisploreguidetop0000brom
I did read “hard drive” and enjoyed it very much, but Sinofsky told me it was not accurate.
Care to elaborate?
Sorry I only recall that someone shared this book on X and I said it’s an interesting book and I love it, and then Sinofsky said the book is inaccurate about early MSFT.
oh that's interesting, didn't know sinofsky disputed it. would love to know which parts he thought were inaccurate
i found it: https://hardcoresoftware.learningbyshipping.com/p/014-execut...
If a reading is popularized due to this website, will you remove it?
Nice collections, cool to have these in one place.
thank you! glad you found it useful.
like your collection thank you for putting this together
thank you - i really appreciate it!
This is very nice. I'm reading "The Tinkerings of Robert Noyce".
Thank you for putting this together. Bell Systems Technical Journal articles would be nice to add.
thank you! glad you're enjoying the noyce piece
bell systems technical journal is a great suggestion. i'll check it out and see what i can add to the collection
update: i've added:
- shannon, a mathematical theory of communication
- shockley, the theory of p-n junctions in semiconductors
- mcilroy, pinson & tague, unix foreword
thanks for the rec!
nice, thank you!
Once I was interested in writing SCP(not sftp) on my own. Good luck finding the spec. I happened upon a random Oracle doc that got me 90% of the way there, and even that felt like a rare treasure.
here's a rare tech book written by the guy who wrote the C compiler for Sun
https://progforperf.github.io/Expert_C_Programming.pdf
looks interesting, thank for sharing!
One of my favorite tech books, but I'm not sure how rare it is.
I'm glad you're excited about what you "built" but it's wildly overproduced as is, and inconsistent in how the animations work. No rhyme or reason to what you've decided to put on the site either, other than "these are some big names that hopefully people will recognize". It reeks of slop.
"Be kind. Don't be snarky. Converse curiously; don't cross-examine. Edit out swipes."
Site guidelines.
I’m sure a throwaway account created at the same time as this comment/post and has only commented on this post really cares about the site guidelines and wasn’t created to boost the content.
fair, i probably got carried away with the animations. will fix that.
as for the books, i was just trying to collect stuff that's hard to find or that most people haven't come across/read. but i can see why the selection feels random. will work on that too.
I think it's neat. I enjoyed giving the plant water. Would be nice to have a list view as an option though, but great idea putting this together.
Agreed. A link to a plaintext listing of titles and authors, would be very useful to those (like me) who enjoy owning physical books.
great idea, this is done.
https://readrare.com/#list
Brilliant, thank you so much!
thank you, glad you liked the plant! i've added a list view.
there's a shelf/list toggle under the headline now, or you can go straight to it: https://readrare.com/#list
Sorry, but if you are developer, slop is coming for job.