modernpink

That all rings true, however the whole industry (or most of it) managed to be built 'on-site' pre-2020. It's not clear where "policing" comes into the discussion here. I actually think the executives calling for more on-site time actually _want_ their workers to take their free lunch, make breakfast and chitchat with their colleagues.

I assume you've not done LISP or LISP-like languages before?

ChatGPT is not the breakthrough it's been made out to be. Yann LeCun has explicated this quite well.

Not sure what relevance this has to tech.

Weighting the access countries by these three factors would be very interesting: Land mass (access to space), GDP per capita (access to human capital), aggregate GDP (access to markets).

Being rich, highly educated, low in criminality, low in illegal emigration means that other nations are more willing to welcome you in. Think of who you would welcome into your own home and they map pretty homomorphically to the same set of rules you would set at the national level i.e. probably someone who won't steal your things, is good company, won't break your rules, or overstay their welcome.

>I spent a month in the lab

Do you mean at your desk? What is a lab in a fintech context?

This is the "economic calculation problem" critique of economic planning proposed by the Austrian school[0]. A planned economy is sort of like having a single sysadmin manage all the resource levers for a data centre, whereas the free market is more like Kubernetes.

https://en.wikipedia.org/wiki/Economic_calculation_problem

The focus in climate change discourse seems to be all about prevention when we've clearly been headed on course for worst case scenario several degree rise the whole time. Emissions have not even peaked yet with China, India and the rest of the third world essentially outdoing all the progress the first world has made over the past decades. Not to mention the coming surge in population of Africa in the coming decades who will all want material lifestyles for themselves.

LDP links to the CIA are barely a secret. Asking for a source there is only to advertise one's ignorance. Not sure about the Moonies part however.

Many (most?) companies do. They make the tradeoff that the false negatives (those from non-targets who would be hired) are worth losing out on in favour of the smaller amount of applications to process because among that smaller pool they have a higher true positive (those from targets who do get hired). In stats/ML speak, they favour high precision [0].

[0] https://en.wikipedia.org/wiki/Precision_and_recall

There are category category errors being committed on all sides. The confusion is arising from the word "type". To some, a "type" of beer is a style of beer characterised by common ingredients, methods and appearance. To others a "type" of beer may be as loosely defined as simply the kind of container it is stored in (can, bottle, draught).

So in which case Korean can said to be a "simpler" language than Chinese (which uses Hanja). But in modern linguistics you would not exactly say any language is "simpler".

> And I fail to see the need for quotes, there's really no denying hangul is simpler than hanja.

Simpler is a culturally conditioned subjective value judgement rather than a scientific one and therefore it expresses an opinionated view. It may be more correct to say it is a more systematic script.

Meatspace Considered Harmful

Do you have peer-reviewed citations to back these points up?

We are headed into a new Cold War that will last decades. The days of globocorps marching into countries in every corner of the world and setting up shop as happened in the 90s-present are now seemingly over. The world will now move away from a trend towards globalisation to one of multipolarity and spheres of influence.

Russia has been caught out by these sanctions, but China looks to have prepared very well with much of its own sovereign digital and financial infrastructure.

Why complain?

Hopefully China gets the same treatment soon. They are a much bigger problem and threat to the West.

Yep, it's really easy to get a job if you know how to speak the language.

The problem is simple, except that Data Scientists seem to resist good software engineering practices and storm ahead to the model development part which leaves a lot of mess for engineers to pick up. MLOps then looks like a 'difficult' problem because of this mess.

We also have to understand Russia's grievances instead of seeing it in terms of an evil side attacking the good side because it's not that simple. The US would hardly be best pleased if Mexico started receiving arms from China and North Korea right at its border.

The Russophobia that has been pouring out is lamentable. And people don't realise this will only push the Russian people further towards Putin to re-establish their pride. And the sanctions are making this even worse. This mirrors Germany's treatment with the Treaty of Versailles leading up to WWII.

I suppose they would think it better to be thought of worse for bad eating habits than it is to be incompetent at abiding by international standards for conducting science, or even to have maliciously released a bioweapon.

Note that this is based on studies by Chinese-only researchers in state-backed institutions [1] that has not yet been peer reviewed. Occam's Razor says that the Wuhan Institute of Virology known for studying coronaviruses must have played a part.

[1] https://www.researchsquare.com/article/rs-1370392/v1

Have wired through £1500.

Interesting that they list CNY but not NTD.

Most CS grads coming onto the engineer market now will have ML exposure through their chosen college courses. As this wave of knowledge makes its way through the industry, the value of ML specialist knowledge will decline, especially as off-the-shelf pre-trained models improve. Very few companies then will be able to justify the luxury of a dedicated in-house data science/ML engineering team. In other words, most software engineers will be ML engineers, in the same way most software engineers today are Docker/cloud proficient engineers.

How much do you hope to sell this to NYT for?

I think a mix of word length and rarity might be the way to go. Information entropy would be a helpful approach here (look up the 3Brown1Blue video on Wordle).

Also, CHOCOLATIER comes up as an invalid word.