Month: January 2021

Amazon logistics

This video covers some, but not all, of the marvel that is Amazon package delivery. Recent innovations include having people on foot / bikes in dense cities to avoid vans getting stuck in traffic, and other things.

2021-03-09: I noticed a few recent improvements.

  • There are now tons of last km delivery people with karts and bicycles
  • They will actually bring the package to your door and ring, solving both the lobby storage and theft issues
  • They’ll take a picture of the delivery if you’re not around.
  • They have an up to the minute tracker on their site.

Can you imagine USPS doing any of these?


2021-11-24: Amazon is now running circles around everyone.

Knopfler says Amazon’s prices were “phenomenal,” $4000 to ship a container from China compared with the $12000 demanded by other freight forwarders. Amazon also simplifies the process since it oversees the shipment from China to its US warehouses. Other services have lots of intermediaries where cargo swaps hands, presenting opportunities for miscommunication and delays. “It’s a 1-stop-shop from Asia to Amazon. It reduces the gray areas where the shipping process might fail.”

2022-07-27: The first Rivian Amazon vans are shipping. The amount of customization is impressive. Logistics is all about little 2% savings here and there.

2022-12-14: Progress in warehouse automation.

The company’s workforce more than doubled during that period, to exceed 1.6m as of early this year. The vast majority of those employees were added in Amazon’s sprawling logistics operation, which delivers packages to e-commerce customers. Amazon has been struggling to manage the size and morale of that group of employees, some of whom have grown restless over the demands of their highly repetitive jobs. The company in October beat back an attempt to unionize a facility in New York state by a nascent labor group.
One of Amazon’s solutions to these issues is robots that could make the roles that many of these workers now occupy obsolete

Orphan Vaccine Carriers

humanity was a lot braver / ingenious 200 years ago. today, people sit on their fat asses debating useless things, like whether human challenge trials are “ethical”:

At the end of the 18th century, smallpox was probably the scariest disease on Earth. It spread alarmingly quickly, and every cm of people’s skin, including their face, would erupt with 1000s upon 1000s of painful, pus-filled sores. Edward Jenner observed something strange, however: People who caught a related disease called cowpox never came down with its deadlier cousin. So in 1796, he began giving people cowpox intentionally, rendering them immune to smallpox and creating the first vaccine.

But the breakthrough introduced another dilemma: How could doctors deliver vaccines to people who needed them? The real trouble started when doctors tried to vaccinate people who were far away. The lymph could lose its potency traveling even the 350km from London to Paris, let alone to the Americas, where it was desperately needed: Smallpox outbreaks there were verging on apocalyptic, killing up to 50% of people who got the virus. Every so often threads of dried lymph did survive an ocean journey—a batch reached Newfoundland in 1800—but the lymph was typically rendered impotent after months at sea. Spain especially struggled to reach its colonies in Central and South America, so in 1803, health officials in the country devised a radical new method for distributing the vaccine abroad: orphan boys.

The plan involved putting 24 Spanish orphans on a ship. Right before they left for the colonies, a doctor would give 2 of them cowpox. After 10 days at sea, the sores on their arms would be nice and ripe. A team of doctors onboard would lance the sores, and scratch the fluid into the arms of 2 more boys. 10 days later, once those boys developed sores, a third pair would receive fluid, and so on.

Corporate donations moderate

Individual donors prefer to support ideologically extreme candidates while access-seeking PACs tend to support more moderate candidates. Thus, institutional changes that limit the availability of money affect the types of candidates who would normally fund-raise from these 2 main sources of campaign funds.

an inconvenient finding for all the fans of grassroots funding. other than a vague “get money out of politics” (how?), i don’t see any policy prescriptions to address this.

Monkey ransom

Shrewd macaques prefer to target items that humans are most likely to exchange for food, such as electronics, rather than objects that tourists care less about, such as hairpins or empty camera bags. Bargaining between a monkey robber, tourist and a temple staff member quite often lasted several minutes. The longest wait before an item was returned was 25 minutes, including 17 minutes of negotiation. For lower-valued items, the monkeys were more likely to conclude successful bartering sessions by accepting a lesser reward.

Better Masks

We need the CDC and the FDA to step up and provide simple, clear, actionable, and specific information that would allow the public to know which masks are reliable and where they can get them, as well as how to upgrade and better wear their existing options.

instead of sitting on their hands running out the clock on vaccine approvals (astrazeneca is still not approved), some mask guidance would be good. instead, everyone at these agencies seems to be playing virtue signaling games to impress their coworkers.

Flying ambulances in NYC

Hatzolah Air, the air division of the volunteer emergency medical service organization pre-ordered 4 air ambulances from Urban Aeronautics. Renderings from press releases announcing the purchase show a futuristic-looking vehicle landing on crowded city streets to assist at the site of an accident.

“Its compact size will enable it to land in the middle of a busy city street, making it a perfect fit for medical evacuation missions by dramatically decreasing the time it takes to arrive on-scene, treat and transport sick or injured patients to appropriate medical facilities,”

while the flying car cosplay on display is a bit premature, there are prototypes, and they seem to at least be thinking about noise.

COVID-19 not over

The new issue is the potential for a really massive spike in Covid cases, where medical systems break down completely. So much of the economic/political status quo seemed to depend on the (largely irrational) optimism that followed the initial vaccine approval. If this gets replaced by a universal “OMG we are so fucked” the damage could be catastrophic–stock market, many industries that have been hanging on the edge, rage at media, doctors, politicians–any “voice of authority”. Even if (rationally) it turns out to be a 2-month issue that the vaccines could have reversed.

Obviously Biden won’t be up for this challenge but no other current or recent politician would have either. I don’t think events will fit any current narratives, and MSM attempts to spin them will fail. The general public won’t specifically blame either Biden or Trump. But if another 10M people are suddenly out of work, no one will understand why it suddenly happened, and no one will trust anyone’s ideas about what to do next.

New variants

Many virologists thought this very unlikely, you could never know that a new variety had higher transmission from mere incidence data: you must understand the biological mechanism. Are they correct? Obviously not.

Why did they think that a new, more transmissible variant of COVID-19 was unlikely? There are several reasons. One, they typically deal with viruses that have been around for a long time, like measles ( > 1000 years). An old virus is going to be pretty well-adapted to to humans. Probably it’s at a local optimum, where small changes would reduce infectivity. But you don’t expect that high degree of optimization in a virus that’s brand new in humans: while spreading to very many people, more than 100M, greatly increases the chance of transmission-increasing mutations. Fisherian acceleration.

Like most biologists and MDs, most virologists don’t know any theory, and in fact don’t _believe_ in theory. For this they occasionally pay a price.

NAT traversal

If the IP address is correct, our only unknown is the port. There’s 65535 possibilities… Could we try all of them? At 100 packets/sec, that’s a worst case of 10 minutes to find the right one. It’s better than nothing, but not great. And it really looks like a port scan (because in fairness, it is), which may anger network intrusion detection software.

We can do much better than that, with the help of the birthday paradox. Rather than open 1 port on the hard side and have the easy side try 65535 possibilities, let’s open, say, 256 ports on the hard side (by having 256 sockets sending to the easy side’s ip:port), and have the easy side probe target ports at random.

If we stick with a fairly modest probing rate of 100 ports/sec, 50% the time we’ll get through in under 2 seconds. And even if we get unlucky, 20 seconds in we’re virtually guaranteed to have found a way in, after probing less than 4% of the total search space.