sfmz

They already have space-based compute in terms of Starlink and Starshield. Is it much more difficult than that ?

We need your salary to buy the machines that will do your job. :|

You spin up agents and want them to paid without opening a bank account... or spin-up hundreds of agents... or your country isn't very well integrated with western banking rails. I think there is more to consider.

An emerging use-case is a bank account for AI agents. I read that Coinbase 'Base' Ethereum layer2 is popular for AI banking.

The unit of work is now a 'prompt'. Imagine recreating different classes of software... there is correlation between complexity (ie # of prompts) and value. Make an app so complex, no one has tried it.

Chess on a sphere, its not as fun as it sounds, but it works. Probably was around $100 of prompting.

https://sphere-chess.replit.app/

"learn to code" is probably over...

Remember the mean-spirited viral meme of 'learn to code' aimed at BuzzFeed journos in 2019? ... not aging well.

I replit'd a travel-site since yesterday: ~8-ish hours: front-end, back-end, auth, mapping, geo-coding, points-of-interest, planning, image carossels, weather-integration (not sure if the weather thing is working), routing.

None of it by hand, no SQL, no Linux provisioning, no learning Map api calls, no typescript, no javascript / webpack bundling, no trying to figure out why i can't get things to vertically center, no css.

There will be like ~1000x more code w/ 1/100x wages.

There's probably space for people with CS PhDs -- but that's not exactly the same as 'learn to code'

Make a yt playlist of lo-fi, there's a lot of it, then u can dl it with yt-dlp.

Mexican lo-fi is a nice sub-genre: https://www.youtube.com/watch?v=_4XHt0iVfRU

Inference cost is dropping 30x this year (see GTC keynote) So imo this is a meaningless blip.

Meta/Google/Anthropic report 75%+ of coding is now AI. For every engineer orchestrating AIs -- X will be let go -- but at what ratio 3:1? 5:1? 10:1? Seems like its at least 3.

Chain a few tools together to make demos, give them away, charge for upkeep.

eg. yt: Nano-Banana + Kling + Claude Code = scrolling-exploded-headphones demo in a few minutes.

https://youtu.be/mhIAd5lVMag?t=133

If you take a year off to orchestrate some complex-workflows with agents... I expect you'll still be hire-able, but idk.

I've never seen a proper defense of XRP, but its been a top 5 coin for like a decade...

In the battle of capital vs labor, the capital wins when the laborers are not needed.

You probably don't need very much news at all. I like learning interesting second-order effects of current events, but they aren't necessary for my life. Here's a few I found interesting lately:

Gen-Z doesn't go to parties anymore because they are on screens all day and/or have social anxiety

Photoshop tutorial business is cratering because people aren't interested in learning how to manipulate photos when they can just create one via text

Japanese guy called do-nothing-guy gets paid to go to restaurants with pretty women in Japan because apparently they can't get a date or are too shy or something terrible is happening with the social fabric in Japan idk what

Its interesting to think about automating the news, you'd need a news feed from every courthouse, twitter feed, sports arena, press release -- run them through various processes, identify trendlines, etc. no small task

The secular papers at the time confirmed the events, so this hot take is low-efefort and low-information and obv wrong.

https://endtimes.video/third-secret-of-fatima/

The generic term is AI-headshot, there are dozens of companies doing this

https://www.betterpic.io/blog/best-ai-headshot-generators

Andrej Karpathy: Let's build GPT: from scratch, in code, spelled out. https://www.youtube.com/watch?v=kCc8FmEb1nY

AI, robotics, drones, self-driving: all poised to take a lot of jobs.

Here's my list of things making it worse: cities designed for cars, no third spaces, poorly designed public space, hostile architecture, no social safety net, replacing social relationships with parasocial relationships, filling every moment of idle time with glacing at your phone, smaller families/smaller extended families, low access to nature, needing a second job (or low free time), too much screentime.

The promise of replacing labor costs with bots is too tempting for Wall Street. The money will keep on rolling in.

The major LLMs should be able to handle that.

Dressing up is a lot more rewarding if you are not obese like 42% of the USA. I think the French still dress nicely is that true?

"The Limiting Factor" had an interesting take: Elon Musk's ability to deploy resources gives Grok a competitive advantage. YT vid goes over the challenges in getting Colossus up (200k GPU) and their deployment of Colossus 2, a 3-gigawatt facility with one million GPUs.

The Limiting Factor: xAI's Plans to Crush OpenAI, Meta, and Google in the AI Race (32m) https://www.youtube.com/watch?v=UoaMveDtCxI

"Wars are won with logisitics" tweet by Elon Musk

Marketing seems way off. Its a personal trainer, not an Ai-Powered-fitness-coach (people are sick of hearing the letters A-I). Your value-prop is to replace $70/session personal trainers and give people ideas of workouts they wouldn't ordinarly think of (get them out of the same routine / rut). I'm reminded of FitDeck from SharkTank -- this guy made bodyweight exercise cards you shuffle to change up your workout each time.

The chat log resides at OpenAi/Meta/Goog. Only they could integrate ads, maybe they will eventually.

Welp, I should have probably told you I'm in a suburb. I'm guessing you are in a cool city. /jealous

What do you do for work? $45k seems low. The low-hanging fruit would be job hop, but tbh I guess the market is pretty bad right now.

Your link is broken. I do know that Mr. Schultz has given lip-service to the fact that Starbucks is no longer a third place, but in the other column I present the fact that there are no longer any comfortable chairs at my nearest Starbucks and the next one by distance is Drive-Thru only.

The two major acheiropoieta are the Shroud of Turin and the Tilma of Juan Diego. They both have a lot of interesting characteristics.

The Shroud

Scientific and Forensic Evidence Photographic Anomalies: The first photograph of the Shroud, taken in 1898, revealed that the negative image on the cloth was a positive image, a concept that would have been impossible for a medieval forger to create [01:51].

Microscopic Details:

Image Formation: According to an optical engineer, the image is formed by tiny fibers one-tenth the size of a human hair and is randomly distributed like dots in a newspaper. Creating this would require technology that does not exist today, such as an "incredibly accurate Atomic laser" [02:47].

No Paint: There is no trace of paint pigment or brush strokes on the Shroud. Some marks only appear under ultraviolet light, which a medieval forger would not have been able to see [02:35].

Blood and Anatomy:

Blood Analysis: Medical experts found human blood of type AB on the Shroud, which contained high levels of bilirubin. This bile pigment would only be present if a deceased man had been severely beaten [03:15].

Flows and Marks: The blood flows on the Shroud are consistent with a man who was crucified, and the wounds match a Roman scourging device called a flagrum and the tip of a Roman spear called a lancea [08:18].

Anatomical Accuracy: The video highlights that the figure on the Shroud shows anatomically correct details, such as the thumb being folded into the palm after a nail pierces the median nerve in the wrist, and that the legs are not broken, consistent with the biblical account [08:50].

Pollen, Dirt, and Other Discoveries Pollen and Cotton:

A Swiss criminologist found 58 types of pollen on the Shroud, with only 17 native to Europe. The remaining pollen came from Palestine and Southern Turkey, indicating the Shroud's historical path [04:16].

Microscopic cotton fibers found in the linen have about eight twists per centimeter, which is consistent with cotton from the Middle East, not Europe [05:07].

Dirt and Flowers: Dirt on the footprint area of the cloth was identified as coming specifically from Jerusalem's Damascus Gate. Images of 28 specific flowers found on the Shroud are all known to grow in or within 12 miles of Jerusalem [06:11].

3D Information: The Shroud contains three-dimensional information, which can be seen with a special machine called a VP-8 analyzer. This is something an artist could not have included in a two-dimensional painting or photograph [06:21].

Coins over the Eyes: The video claims that a Jewish custom of placing coins over the eyes of the deceased is evident on the Shroud. A coin-like image over the right eye has 24 coincidences with a lepton coin minted by Pontius Pilate between 29 and 36 A.D. [06:46].

Proof that the Shroud of Turin is the Burial Cloth of Jesus Christ! https://www.youtube.com/watch?v=WRB16BARvz0

The Tilma of Juan Diego

Miraculous Features of the Image

Lack of Underdrawing: Infrared photography has shown that there is no under-sketching on the image, which is considered inconceivable for a 16th-century artist to create such a detailed portrait without one [00:46].

Durability: The tilma is made of a rough cactus fiber that typically disintegrates in 15 to 30 years. However, the image has remained for over 481 years without any fading or cracking, even after being exposed to candle smoke [01:34].

Chemical and Physical Resilience:

In 1778, a significant amount of nitric acid was spilled on the image, but it only caused slight stains that are still visible in the upper right corner [03:14].

In 1921, a bomb hidden in flowers exploded beneath the image, damaging a crucifix but leaving the glass protecting the image completely unharmed [03:42].

Astronomical and Geographical Correspondence:

The stars on the mantle of the image correspond precisely to the constellations in the winter sky on December 12, 1531, as if viewed from outside the heavens [04:19].

The flowers on the tunic align with the locations of volcanoes in Mexico [04:45].

Additionally, the constellations Corona Borealis, Virgo, and Leo would appear on the image as a crown, over her heart, and over her womb, respectively, if her tunic had been filled with stars [05:50].

The Eyes: The eyes of the Virgin's image contain microscopic reflections of a bearded man, which experts believe is Juan Diego. These reflections follow the Perkinje-Sanson law, which governs how reflections appear in a human eye [08:06]. A computer engineer also discovered what appear to be multiple figures, including a whole room of individuals, within the eyes [11:00].

Changing Appearance: The video mentions that the Virgin's face appears to be Native American from one distance and of lighter, European descent from another, symbolizing the unity of the two peoples [13:10]. This effect is described as being similar to the coloration found on butterfly scales or beetle wings [14:06].

Constant Temperature: The image consistently maintains a temperature of 36.5°C (98.6°F), which is normal human body temperature, regardless of the surrounding environment [14:44].

The Amazing And Miraculous Image Of Our Lady Of Guadalupe https://www.youtube.com/watch?v=Ds7nD_QNeKA