Shrezzing

While it's accurate that task complexity is the shoreline paradox, it also has the same obvious solution as the shoreline paradox in day-to-day scenarios: Agree a resolution ahead of time and discuss the measurement in those terms.

Nobody wants to know to-the-second when a project's Jira board will be complete, in the same way nobody wants to know to the atom-width how long the shoreline of Britain is. They want to know if the project is going to finish it in-or-near the last week of Q3, they want to know how many meters the shoreline is.

The MS documentation for Excel is a pretty good resource to learn from these days, with text descriptions for just about every function, and videos for most too. This page [1] is an especially great resource. It's got a "top-10" list, which is a good place to start, as it covers the majority of things you'll see in a normal business document.

After those 10, they break them down by category, and have one for Financial Functions, which is going to be useful for you. Similarly, the Logical, Math & Trig, and Stats functions, will all be useful for looking at a Finance firm's spreadsheets.

[1] https://support.microsoft.com/en-gb/office/excel-functions-b...

I fully agree with the point you're making, but to be pedantic, clocks and timezones are themselves assistive technologies. The precursor to them was taking a look at your shadow and making an educated guess.

If a GitHub user fails to respond to a DMCA request, and GitHub doesn't remove the content in question, GitHub as a service provider would then become party to the complaint under the copyright safe-harbor rules (a suit that GitHub would definitely lose, and can have fairly drastic downstream consequences for their status as a safe-harbor going forward).

It's not that users are irrationally responding to DMCA requests when they don't need to, its that as a service provider, GitHub defaults to a stance where a claimant's DMCA is automatically processed if a repo author doesn't reply within 1 working day. This effectively means that for DMCA requests of all types on GitHub, "no response" is synonymous with "I'm guilty of infringement, plaese take down the content".

A DMCA claim response in GitHub is admissible as testimony should the case ever make it to court, and therefore carries the penalty of perjury (both parties are informed of this before opening/responding to the claim in the GitHub UI). So unless the repo author is absolutely certain that they're not hosting DRM circumvention software, they have little-to-no recourse but to allow the takedown request to go ahead.

If someone says they're going to sue you, and it's black-and-white that they'd win the suit, then it's pragmatic for all parties concerned to just skip the time & cost of a court appearance, and just do the thing the court would otherwise force you to do.

In a sane world, civil suits between firms should really only go ahead if the case's result is ambiguous, and the courts in this case would almost certainly side with the takedown request - and moreover would be pissed at Github for wasting the court's time on it.

$9tn - around 2.1% of global wealth. Equivalent to 4x more than the physical USD$ currency in circulation, and equivalent to around 2x the value held in deposits at commercial US banks.

YCombinator might be the literal antithesis to the suggestion that $600k isn't a meaningful sum of money in the face of large problems.

We could reasonably plant ~1tn trees without encroaching on current food production levels (although we'd have a fairly catastrophic impact on biodiversity). This would sequester around 100Gt of CO2. So far, we've emitted about 2800Gt of CO2, with around 950Gt remaining in the atmosphere (the rest is already in trees/oceans etc).

So at the moment we could sequester all of the excess CO2 that's still up there in the atmosphere, in exchange for significantly altering the natural world.

However, planting 1tn trees takes time, and trees take a very long time to grow and sequester carbon. Once they're fully matured, they stop sequestration. While they're growing, humans would continue emitting ~35Gt CO2 per year at current levels, with "business as usual" projections of ~100Gt per year in 2100 (in the word "4 degrees of warming" scenario).

So although planting lots of trees is a valuable tool in the carbon sequestration arsenal, we really need fairly radical decarbonisation alongside it, if it's to be anything more than buying a decade at the end of the century.

> The technological roots of LLMs go back many years. Yet, today’s experience looks like more than the continuation of a preexisting trend. Something in the zeitgeist changed recently,

I think the thing in the zeitgeist that changed isn't anything to do with the technology, it's that truth itself is tangible in the modern society. We're still grappling with the "post truth" era. We're all constantly deceived by marketing, PR, politicians, adverts, social media, each other, etc - and we're all fully aware of that deceit.

In the face of that, an LLM that's only accurate 90% of the time looks acceptable, like a complete technology, and therefore worthy of investing in.

So... Indentured slavery?

SpaceX is benefiting enormously from their supply chain receiving a $50bn injection from the Artemis program.

I think the downtime associated with other services could just be people choosing alternative sites for their social media time.

Facebook + Insta makes up a huge share of the social media market, and when they go offline, it'd be natural for their competitors to receive large sudden upticks in trafic they're not immediately prepared for on a Tuesday morning.

What I thought this product was:

> a WAF for your website that detects if the site's being scraped for content by an AI. Defending against AI

What this product actually is:

> A WAF for your chatbot. Defending the AI itself

The net after-tax salary in France is €2,450, and in Germany it's €2,750, compared to the €850/mo this kid spends.

I can't see how you could survive long-term on around 1/3 of a nation's average salary, even in the most rural of villages with hardly any amenities, rent alone is going to take up the majority of your budget.

You could probably get by reasonably well in Greece, and could live very well in the Balkans.

The US Money Supply (M2 & M1) has been falling pretty consistently since Russia's full-scale invasion of Ukraine in 2022. M0 is also down overall since the invasion, though it's gone through ups and downs.

I'm not certain the statement "Inflating the money supply by billions of dollars to fund several expensive conflicts" reflects reality.

>Your site should be completely usable without JavaScript... If your site is not usable without JavaScript, I am not coming back in any way

The search page on the site doesn't work with js disabled.

That's fair, and you're entirely correct that its nobody's business but his own.

This is somewhat off-topic, but does Wolfram have a degenerative muscle illness? He puts his right hand down at around 30 minutes, and then doesn't move it again for hours. That and his unmoving legs are strikingly reminiscent of a friend who had MS.

I'd argue that this article could be read more like "start as a monolith, and then move towards microservices when sensible", except the penny hasn't fully dropped for the author that the sensible time for their organisation is right now.

The company appears dogmatically locked into their idling python code, their SQL making unpredictable commits, and their SQL framework making it difficult to make schema migrations.

This is a financial services company describing data-integrity bugs in their production platforms.

Oxford Economics estimates that there's a 6.7x multiplier on each new role created in the semiconductor industry[1]. That is, each role in the semiconductor industry supports an additional 5.7 jobs indirectly in other domestic industries. The only two sectors that are higher are Computer Storage manufacturing and Aircraft Manufacturing.

You'd need to split the pay between all 6.7 newly created roles, meaning they only need ~$60k take-home or so each - which is right around the US mean salary.

[1] https://www.oxfordeconomics.com/resource/chipping-in-the-us-...

Padding's great on my desktop with two 24" monitors. On my one-screen 13" laptop it's less welcome.

Where did you find the data source for the USA?

It'd be amazing to see this data mapped across time, as it'd be a relatively good indicator of grocery inflation.

Depends a bit on the branch/focus of economics you're looking at, and which nation/region you want to read about, but a broadly uncontroversial list would be:

The Quarterly Journal of Economics, Oxford University Press (colloquial known as "the QUGE")

Econometrica, Econometric Society

Journal of Economic Literature, American Economic Association

Journal of Economic Growth, Springer

Journal of Financial Economics, Elsevier

Maybe swap in the Journal of Political Economy if you're looking for more political leaning research. Swap in the American Economic Review if you're interested in more American-centric research.

In the same theme, something to add to their list of common passwords is the What3Words database of locations

https://what3words.com

It's something like 50trillion sets of looks-random strings. That's quite a lot, but if the list could be narrowed very significantly to get some likely results by selecting locations in:

1) cities where a company is physically located

2) large capital & global cities

3) significant landmarks

I see sysadmins using the tool all the time as a temporary password generator.

>So 0.0012% of CO2 in atmosphere caused by humans. Nature is responsible for the remaining 0.0388% in atmosphere.

Unless you're discussing literal "human expelled co2", as in "the measure of peoples exhalations during breathing", then you're out by several orders of magnitude. The pre-industrial co2 concentration in the atmosphere trends between 180ppm - 300ppm. In the last hundred years, the concentration has increased to 420ppm.

https://climate.nasa.gov/vital-signs/carbon-dioxide/

For reference, each 1ppm in the atmosphere is around 7.8 gigatonnes of CO2. Humans currently create around 40gigatonnes of CO2 per year, roughly equating to 5ppm added per year. We add equivalent to 0.8% of the atmospheric CO2 per year.

On longevity+userbase I think Minecraft and Roblox might have both beat. Both are 15+ years old, with enormous userbases, crucially their userbases include huge numbers of kids.

WoW and Counterstrike have large loyal fanbases, but I'd be surprised if the age of the average user didn't increase by around 1 year per year.

>“Tell me about you. If your life was a book, give me the chapter titles from your birth till now.” Once you’ve gotten the overview, dive into each “chapter” and plumb the depths for their real stories. Go back to their childhood! I learn a lot about their grit and commitment to excellence from their basketball obsession or maybe their experience caring for a dying parent.

This paragraph starts bad and somehow gets worse until a spectacular crescendo of unapologetically shamefully exploitative nonsense.

That it's posted in an article about techniques for interviewing 19-21 year olds makes it so much more insidious. The article in its entirity reads like a sociopath's guide to bullying teenagers.

> I realise this is a silly thought experiment but the point of comparison should be, at the very most, the human industry generated CO2, not all natural CO2

The pre-industrial levels of CO2 were around 280ppm, while the current CO2 levels are around 440ppm.

For Cl2, in occupational-health situations the permissible exposure is around 1ppm per hour, or 3ppm for 15 minutes. So the 160ppm we'd throw into the atmosphere in this thought experiment (or even the 52ppm at the suggested discount) is going to cause some pretty big problems pretty rapidly.

There's a table on the 2nd page of the paper that gives a summary of what goes on when the VM's container is initialised the first time:

INITIALIZATION PHASE: (1) Start the VM.

(2) Fetch the container image and the function's code.

(3) Initialize the container.

(4) Start the language runtime such as Python or PHP.

(5) Load common libraries into memory.

(6) Load the function code into memory.

(7) Optionally, do JIT compilation. INVOCATION PHASE:

(8) Invoke the function multiple times as needed. SHUTDOWN PHASE:

(9) Stop the container if it receives no requests for X minutes (X=10/20/10 minutes for AWS/Azure/OpenWhisk respectively).

(10) Optionally, stop the VM

>I think this needs to be more specific; what resources are exactly locked to a VM?

Steps 1-7 are time-consuming processes, which load a container into memory so that step 8 can run rapidly when invoked later. Step 8 can be invoked an unlimited number of times on a VM once steps 1-7 have completed. Only the first Function execution on a VM pays the startup penalty.

>Are they pinning VMs to specific CPUs or something that prevents the host from scheduling other tasks there? If so, why?

They're pinning memory to VMs in their FaaS infrastructure, which in turn is memory consumed on their physical servers. A VM uses some number of MBs of memory to host a Function's unique container, its runtime, and its executables. Meta's FaaS solution completes several trillion operations per day, which is in the tens of millions per second avg. This relatively small amount of memory per VM becomes enormous at tens-of-millions-per-second scale.

>If so, why?

The resources are pinned to the memory by the first function initialisation, so that if there's a second request for the same Function, it can be executed on a pre-warmed VM, without going through the time-consuming steps 1-7. A trade-off between memory-consumption and time, both cost money.

There is some optimal amount of time between the end of the final Function invocation of step 8, and the start of step 9. If the amount of time is 0-seconds, then every Function request goes through steps 1-7. While if the amount of time is 1-hour, at midnight when Meta's large one-a-day batch jobs run, most of the servers fill with idle VMs until 1am, waiting to satisfy a batch-job that never instantiates.

Given that infrastructure cannot be aware of which Function execution is "the final Function invocation", it's challenging to pin this amount of time down. In a public cloud, that optimal amount of time is set in the 10s of minutes. In Meta's case, as a private cloud provider to themselves, they have a unique awareness to the usage pattern of their Functions, and therefore can predict more acurately that they don't need 10-minute-idle-Functions. They've instead been able to reduce the idle time to 1 minute. This releases Containers & VMs, and returns memory back into the server pool.

>do you have some additional sources you could point me to, or are you also basing all this on this one paper?

Everything I've said about Meta is based on this paper, because it comprehensively describes Meta's unique circumstances, and how they approached the problem. The paper itself has dozens of additional sources.

The server in this context is a "warmed" Function VM. The first time a VM instantiates, it goes through all of the setup required to run a function. That includes generic reusable stuff like Operating System setup, but also Function-specific things like JIT compilation of the Function software, library loading, language loading, all of which are specific to the Function being executed.

The Function VM is hyper-optomised to that one Function's task. There's compute-time costs in changing its role from Function A to Function B. In the public cloud, the only cost you save on is starting up the OS, if both Functions happen to use the same OS. It's faster in most public FaaS infrastructure settings to just destroy a Function VM than it is to transfer a VM from Function A's configuration into Function B's configuration. Similarly, it's faster to just leave the VM running for a little while, just in case the Function is invoked again.

When the paper discusses VMs idling, that's the scenario they describe, where servers (or more accurately software VMs inside of physical servers), are locked into a specific function for some number of minutes. These minutes are dead computing time.

Meta, as a private cloud provider to themselves, can do some interesting things to reduce the cold-start of their Functions, meaning they can run more efficient clouds if they choose, by massively reducing the idle-time that's an industry standard in the public cloud.