Closi

35 ms·

421.
by Closi 3y ago
This is probably true - i.e. you could make an even smaller model and then likely come up with an even-simpler explanation for how it worked.
422.
by Closi 3y ago
I think you are going the wrong way - I would start with the best possible AI (which will probably be GPT4) and see if it can do it, and then walk backwards to local AI deployments (which are currently significantly weaker).
423.
by Closi 3y ago
> They would have different tax rates based on agreements between countries. Yes - but the issue is how to get global agreement between all countries. i.e. Will the USA agree to be taxed substantially more per unit than China and Russia?
424.
by Closi 3y ago
Well China still sees itself as a developing country, and would make up about 30% of CO2, so it all depends on definitions and international agreement.
425.
by Closi 3y ago
That's only one dimension - Another major dimension is country. The same tax applied globally will have very different effects in a third world country to a first world country. Farmers in the USA will react very differently to a 5p pe
426.
by Closi 3y ago
The actual problem is a little more difficult - clearly fuel is often taxed, but the question is how much tax should be paid on which fuels, in which industries and in which countries. Should an American putting fuel into their hummer pay t
427.
by Closi 3y ago
It's absolutely solvable - and presumably this is the sort of functionality that more legal-specific AI applications offer (probably mixed with some sort of knowledge graph that contains facts). Seems like the issue here was that a law
428.
by Closi 3y ago
Well drug testing for employment usually will be similar to a field test in the first stage, but at failure will then get sent to a lab for proper analysis (at least in my experience). Laboratory analysis should (obviously) be required for
429.
by Closi 3y ago
Sure, but probable cause probably can't be justified if you are drug screening everyone attending a political rally - that would probably be easier to argue as abuse than a stop of a single person on the street.
430.
by Closi 3y ago
That sounds like a problem with arrest records across the board then, rather than a problem with drug testing.
431.
by Closi 3y ago
Depends - I would say 4% is probably accurate enough evidence to support an arrest (i.e. they are detained someone in order to investigate a crime they are suspected of committing), but not enough to support a conviction. Arrest != convicti
432.
by Closi 3y ago
Good point - so I guess it could be anything from 0% to 40%, although I guess my broader point is that the measure of absolute percentage points could hide significant ‘actual’ growth.
433.
by Closi 3y ago
Gaining 1% of market share from a baseline of 2.4% would mean growing their share of the market by 41%
434.
by Closi 3y ago
Having an occasional Indian takeout isn't living it up, but having $120 Indian takeout every night and spending $43,800 a year on dinner is. We will all have different definitions of living it up, but I would say you can live a good fo
435.
by Closi 3y ago
It's easily into the 'diminishing returns' category in terms of lifestyle. You can buy a lovely house, support a family, 80k is still a full salary for most people and will certainly support 'standard day-to-day' li
436.
by Closi 3y ago
I'm not talking about a language that an LLM outputs. I'm talking about a higher-level language that outputs Javascript, Java, C# etc (but presumably some sort of lower level byte-code in the future). A language where intent and c
437.
by Closi 3y ago
Well, I guess I'm really talking about the price of the service rather than the cost to run it really (i.e. what would be a competitive price for the service considering other offerings), but appreciate that the price has to be less th
438.
by Closi 3y ago
Yes, I think the current state of technology (just starting) is a technology that complements developers. Although my personal belief is this is more likely going to be an improvement similar to Assembly -> C (even if the LLM component d
439.
by Closi 3y ago
I'd pay for YouTube Premium if it was un-bundled from Youtube Music to make it cheaper. No reason YouTube Premium should be more expensive than Spotify or Netflix. As an additional note - I have previously subscribed to YouTube premium
440.
by Closi 3y ago
Why are your tests changing multiple variables? A simpler test would be keeping the development environment the same and then adding GPT4 to see if there is a statistically significant and meaningful speed increase (of a decent magnitude).
441.
by Closi 3y ago
Fixing bugs in code it writes itself is very different to diagnosing (i.e. identifying, not fixing) bugs in my own code. When stuck I often paste code into ChatGPT and ask why it doesn't work, and it will often help me quickly identify
442.
by Closi 3y ago
> I think it's as close to the perfect benchmark as you can get. Depends what you are benchmarking for... If you are benchmarking the ability of the solution to solve LEETCODE challenges, that is different to the ability of GPT4 to
443.
by Closi 3y ago
This is amazing! Didn't realise parts of the code was available. If anyone was in any doubt, VBA6 does support: number = -number
444.
by Closi 3y ago
Well it could be every few miliseconds. Sometimes these sort of screwdrivers can drive other processes (e.g. once it confirms that the screw is correctly affixed, send a message to the manufacturing system to tick off that 'task'
445.
by Closi 3y ago
Because they fall out the sky and occasionally their windows fly off.
446.
by Closi 3y ago
Super impressive app by the look of things, but as you asked for feedback, it is (to me) very confusing on the product side of things (i.e. what is it and why does it matter to me). As a business user it's not clear how I would use it,
447.
by Closi 3y ago
> Also, what is the point of giving this service your AWS creds rather than them using AWS directly? Since they charge you anyway? Presumably so that if/when the service gets abused and Amazon decides to ban the offending account, i
448.
by Closi 3y ago
All jobs are transactional, the transaction is just more clear-cut with #1.
449.
by Closi 3y ago
I disagree - It depends on your career aspirations, your skillset, the company, and the type of software you like making. If you are a great communicator and have a great ability to just 'get stuff done', you might be a superstar
450.
by Closi 3y ago
It will depend on what you are looking to get out of advertising - immediate sales or engagement/awareness. If you are an eCommerce business selling a widget then you can easily get a cost per conversion, and then it's pretty easy

More ›