primitivesuave

I learned how to use embeddings by building semantic search for the Bhagavad Gita. I simply saved the embeddings for all 700 verses into a big file which is stored in a Lambda function, and compared against incoming queries with a single query to OpenAI's embedding endpoint.

Shameless plug in case anyone wants to test it out - https://gita.pub

They do make some effort to push the “confidential cloud” aspect of their product, where the AI is supposed to somehow operate on your encrypted data, but these days I only believe it when I see the white paper. It is unfortunately quite common in our industry is to posture as more security-conscious and privacy-focused than you actually are.

At the end of the article it points out that a foreign toy company was successfully sued for making a puzzle from Leonardo da Vinci's Vitruvian Man. I'm curious what enforcement authority the Italians have over a German company, if any. The court ruling is effectively saying that the historical IP of the Italian peninsula belongs to whichever Italian museum happens to have the original artifact in their possession, but would a German or EU court uphold this?

This is incredible. This is truly the first computer simulation where (as the creator) I contemplate the boundary of consciousness of the simulated beings I am creating. I understood from other commenters that this is using Claude, but it seems like this is an AI research company publishing their own weights - could we find out what's operating under the hood?

This is a great example of an open-core business model that I hope others will follow. Make something useful for everyone, and sell additional features in a hosted version (grit.io). I actually didn't even know there was a hosted version until I clicked the outgoing links from the documentation.

I've experienced how much time these migrations take (and how error prone they are), especially once you have double-digit team sizes, so I definitely see the value proposition.

The headquarters are in Champaign, IL which is a small town with a much lower cost of living than Silicon Valley - so salaries are 20-40% lower than Automattic or other comparable mid-sized firm. The main revenue driver is Mathematica (now called the Wolfram Language) which has several lucrative and long-standing academic licensing deals in place. I don't think the other endeavors (Wolfram Alpha, Wolfram cloud, book publishing) have been nearly as profitable.

(I worked at Wolfram Research around a decade ago, when it was ~500 employees)

Keep in mind that zero Americans were actually surveyed to arrive at this figure. The methodology was simply extrapolate from price indexes of things we think people pay for.

https://www.onemainfinancial.com/personal-loans/where-americ...

That's true, but also consider how the entire world paid attention to the whereabouts of a missing submersible when there were a couple billionaires on board (OceanGate). I doubt the traffickers would want to risk that kind of exposure ;)

This actually seems to be happening already in regards to manufacturing. I read an obscure article a while back about autonomous narco submarines surfacing in Spain, which suggests a certain sophistication and standardization of the manufacturing process [1].

> The [d]rone submarines are essentially identical, suggesting series production of a refined design.

They also found a "narco surfboard" which presumably disguises drug shipments in an abandoned surfboard. In the arms race between traffickers and enforcers, I'd imagine both sides will be continuously developing increasingly sophisticated technology for underwater smuggling, especially given how much money is involved - so I could totally imagine a high-tech manufacturing industry emerge as a byproduct. I would also find it quite funny :)

1. http://www.hisutton.com/Narco-Drone-Submarine-Found-In-Spain...

Check out narco submarines - they have people in the jungles of Columbia building submarines that they pack with tons of cocaine (literally), which are discarded after a single ocean crossing. In general, it seems to be more economical to ship illicit materials with cheap/disposable vessels and vulnerable/impoverished people, rather than risking an expensive yacht to government seizure.

The PM at the time took contention to the publication of that article, saying that there was no increase in actual habitable land.

https://web.archive.org/web/20190326140508/https://kmt.news/...

Absolutely - I created the website pppwatch.com, which currently has searchable aggregates of the data from the official SBA data source [1]. The individual loan search is down right now, not because of any C&D letters, but simply to save some money on the search cluster while I have been revamping the site in my free time with some new analysis.

1. https://www.sba.gov/funding-programs/loans/covid-19-relief-o...

I got multiple C&Ds for publishing public information, provided directly from the US government as giant CSV files (PPP loan data, provided under FOIA), through a web interface that made it easy to search loan recipients and analyze distribution of COVID relief funds.

Many of the C&Ds came from people who were later indicted for defrauding the government, some of them in hilariously inept ways.

To add some context here - the concept of a unified "Germany" has only existed four times in history. First under Charlemagne in the 800s, then under Otto von Bismarck in 1871, then under Adolf Hitler until the fall of the Third Reich in 1945 (assuming one considers annexation of the Sudetenland and the Anschluss to fully encompass Germanic people), and finally the modern state of Germany since 1990.

Bootstrapping requires a rare breed of founder who is financially secure enough (or frugal enough) to manage life on a reduced salary, capable of building useful products on a budget, and willing to risk the missed opportunity cost of a cushy Big Co position. In my experience, the hardest thing to master is the actual building - plenty of people are bootstrapping "theoretically useful" inventions in basements/garages/etc, that don't solve actual problems. Such people can really benefit from VC oversight and business acumen.

This is unfortunate. My favorite feature of EveryNoise is the geographic categorization of music genres:

- https://everynoise.com/cities.html

- https://everynoise.com/countries.html

There are also features for "hyper local" discovery, i.e. finding the individual tracks that are most popular in your city. For music lovers, this site provides the initial breadcrumb for many fascinating rabbit holes.

- https://everynoise.com/hyperspace_house_concerts.cgi

The assessment of Indian history is quite accurate, though it is important to note the distinction between the Greek/Chinese written history and the Indian tradition of oral history. Foundational works like the Mahabharata and the Bhagavad Gita were transmitted orally, from teacher to student, over the course of many centuries. Various memorization techniques (e.g. the eleven pathas of Vedic chanting [1]) were employed to promote perfect transmission of information.

This presumably means that during the many centuries of internal strife in the Indian subcontinent, warfare and disease could erase the history of the vanquished by simply eliminating the learned few who had memorized it.

It's also worth pointing out that contemporary Indian history is extremely biased. In the museum adjoining the Golden Temple in Amritsar, the historical account of the 1984 Sikh separatist movement universally refer to the event that led to the assassination of Indira Gandhi as the "Blue Star massacre". The Partition Museum, located less than a mile away, uses the generally-accepted name of "Operation Blue Star", but is riddled with factual inaccuracies and anti-colonial editorialization in many of its displays. The underlying motives of many Indian historians is to rewrite it to suit their present purposes, which is why P. N. Oak [2] has such a huge following in the highest levels of Indian government.

My nieces and nephews in India are taught "Indian exceptionalism" in school. Many Indians will erroneously claim that Western ideas like germ theory are either appropriated from, or inferior to, the ancient practice of Ayurveda. And even when it comes to what little historical information one might find in the Mahabharata, the words are treated as mystical sounds rather than something to be analyzed and understood.

1. https://en.wikipedia.org/wiki/Vedic_chant

2. https://en.wikipedia.org/wiki/P._N._Oak

I highly recommend watching The Wire as it provides a stunningly accurate depiction of how this strategy failed in Baltimore. As a sibling comment points out, the "war on drugs" was an abject failure, because the people getting imprisoned for their role in the drug economy were usually street-level dealers and foot soldiers. The people who run criminal enterprises tend to insulate themselves from its everyday operations, which is why they could only get Al Capone off the streets by charging him with tax evasion.

I wholeheartedly agree with your notions on drug usage and decriminalization of certain substances though! You might be interested in Dr. Timothy Leary, who wrote extensively on "control over your own nervous system" as long as it did not affect others [1] (on page 49).

1. https://repository.arizona.edu/bitstream/handle/10150/565255...

I created this Scratch proof-of-concept for 3D pong (https://scratch.mit.edu/projects/1881311/) over a decade ago when I used to teach kids. I also found several other examples on the Scratch website which show how to build 3D graphics on a 2D canvas. After many years of teaching kids to code, I would say the best introduction for a 9 year old is to start with block-based environments to build structural thinking fundamentals, before introducing him to typed programming languages.

The issue I see in starting in Unity or an equivalent high-level abstraction (e.g. threejs.org), is that it doesn't build strong fundamentals in how you think about solving problems. Many kids these days think they are "coding" when they are really just modifying a collection of examples and starting points (which has fueled the burgeoning industry of summer camps teaching "Minecraft modding"). As is the case for many other skills, great amateurs make great professionals.

Virtually every homeless shelter will refuse to house someone who is drunk, on drugs, or belligerent. This article only makes a passing reference to people who refuse shelter so they can continue their substance abuse, but this is a salient reason for homelessness given by many people on the streets (especially in SF).

This move by Apple is likely to serve as a deterrent for other technology companies thinking of ways to minimize Apple's commission on in-app purchases. When some business analyst points out how many millions the company might save by shifting payments away from Apple's network, the legal analyst will nip it in the bud by citing this example of potential litigation costs.

You're right, I didn't put that together, and now find myself descending down the rabbit hole of Navy slang :)

https://en.wiktionary.org/wiki/Appendix:Glossary_of_U.S._Nav...

The primary sources are hard to find on this, but the common year of first usage in English (1914) seems to coincide with the first sinking of a ship by a German submarine-launched torpedo [1].

1. https://en.wikipedia.org/wiki/HMS_Pathfinder_(1904)

I wholeheartedly agree as well. The only nitpick I'd have on this incredible dissertation is on the "U" association with "U-boat" - this is an anglicized form of the German word "Unterseeboot" (undersea boat) which was itself shortened to "U-boot" (presumably the etymology of the Swedish "ubåt"). The first known use of the term "U-boat" is 1914, while the Germans and Swedish armed forces were commissioning submarines around a decade prior (in 1906 and 1904 respectively).

The CIA probably realizes it doesn't need to fund rebel groups [1] or use local journalists/clergy [2] to instigate a regime change these days - they can just flood a target country's social media with AI-generated propaganda.

1. https://en.wikipedia.org/wiki/Iran%E2%80%93Contra_affair

2. https://www.intelligence.senate.gov/sites/default/files/hear...

I really appreciate you sharing these candid insights. Let me tell you (after over a decade of deploying cloud services), some rogue user will always figure out how to throw an unforeseen wrench into your system as the service gets more popular. Even worse than an outage is when someone figures out how to explode your cloud computing costs :)

I completely agree with your point about Pigafetta only writing from his own perspective. Perhaps a more poignant example was his naming of modern-day Guam as “Isla de Los Robos” (island of thieves) because he could only perceive their communal society as one where people are constantly stealing from each other, despite spending over a month living amongst them while giving accounts of their harmonious society.

As for the general state of academia, I would opine that modern scholarship’s “dark mass” is rooted in postmodernism, and the most successful way to operate as a modern academic is to present all evidence from a racial/colonial perspective. Although your guess is as good as mine, as is the BBC’s, since people also need to contend with the potential Internet backlash from presenting their views nowadays.

I can completely relate to this, and was in a similar place not too long ago. My personal recommendation would be to check out Vipassana (dhamma.org) - it was a life-changing experience, and the first to actually address some of the mental issues I was facing.

This is exactly my point - the "woke" perspective is that every scholar of that era is to be denounced as a white supremacists, and their work ignored, without any examination of the literature or citing of a single specific example.

I am not saying you're incorrect - for example, the German Anthropologie school (largely centered around the Kaiser Wilhelm Institute from 1927 onwards) was famous for its pseudo-scholarship on racial disparities. However, this is by no means a blanket statement on European scholarship - one notable exception that comes to mind is Max Müller [1], who was famously accused of being anti-Christian for his admiration of Eastern cultures, and was "deeply saddened by the fact that these classifications later came to be expressed in racist terms".

Likewise, there are several German authors whose books sit on my parent's bookshelf that provided objective anthropological history of India (e.g. on the Kuru kingdom and PGW civilization) without editorializing a white supremacist narrative onto it.

Also, the Nazi party only came to prominence in the 1930s, after they had sufficiently terrorized the population to go along with their racial policies, and forced academia to go along with their unscientific ideology - a notoriously similar trend is going on in America today.

1. https://en.wikipedia.org/wiki/Max_M%C3%BCller

Gladiatorial combat was just a drop in the bucket. A single Roman mining operation in Spain [1] had tens of thousands of slaves working to death. Even the famous slave revolt of Spartacus ended with the spectacle of 6000 slaves crucified along the Appian Way [2].

1. https://en.wikipedia.org/wiki/Las_M%C3%A9dulas

2. https://en.wikipedia.org/wiki/Appian_Way#The_crucifixion_of_...