sackfield

It is reasonable to break PRs up into smaller chunks, but there is a limit. There are frequently reviewers who get very zealous about this and insist on breaking things up beyond what is reasonable, for example if breaking it up would destroy the intent, or if the "thousand line" PR just contains lots of tests (AI's love to write tests, and I love that they do that). Some tasks are just long, and its important to contextualise this when reviewing.

In the end though, these reviewers will die off like the dinosaurs. The article actually states that they find the idea of reviewing a large PR with AI bad because "it wastes your tokens reviewing a reingesting code that was already made by an AI". This doesn't make a whole lot of sense, AI will frequently reingest AI generated content, evals are a great example of this.

Just after this the article touches on the real issue at play: "okay, great, why did you put it up for a human review at all then?". Indeed, this is a good question to ask, why do we put it up for human review? I would wager that they don't actually want human feedback, a human has placed themselves as a gatekeeper and thus must be placated, and probably chooses the most inefficient way to keep that gate slowing everyone down who has kept up with the technology of our times.

Wow I thought I was the only one. I have also had these experiences, including abused by people on this site for the crime of saying I miss the office.

I am back at the office now, and am enjoying the separation between work and home.

One thing I wonder, if you look at the comments here it is clear the moderation and the general hostility of the site was a big reason why a lot of people left or didn't engage with it. Yet, I can't recall SO doing anything about this, and its confusing to me why. Was it:

a) They don't acknowledge this was a problem.

b) They are unwilling to tackle this problem.

c) They are unable to tackle this problem.

I'd love to know what the people who run the site (from a business perspective) think about this. It might offer some insight into its cousins that are going down the same path (reddit / wikipedia).

I've always found it quite sad and cringeworthy when people talk about AI's water usage. The first thought that comes to my head is whether its even worth trying to talk the person out of their delusion, or just accept that they are lost and can't be helped.

How is it going in the other countries that are trying this? Do we know yet? If not, why move forward with legislation based on untested theories, especially when other countries are currently testing it for you?

I'm not interested in defending the trend as much as I am in understanding it, which the article promises but doesn't deliver.

To begin to understand this, you have to examine the actions of the heroes as critically as the villains as defined in this piece. The article lays out clearly the negative desires of the villains, and the positive desires of the heroes, but do the heroes have any negative aspects? Does the EU simply want to protect consumers or is there an argument that they are the law to unfairly targeting American companies? What about the villains, do they have any positive aspects? Does Musk want humanity to keep existing to the point where he is willing to put capital on the line to give our species a backup planet?

The point of this comment isn't to defend the villains or vilify the heroes, its to recognise that these issues are not simple as defined by the article, and in presenting them as simple you don't end up with an understanding of the core question: "How the Tech World Turned Evil".

If you are actually looking for an article on "How the Tech World Turned Evil", you are going to be sorely disappointed.

This article is, as you might expect, the usual cast of villains and the usual cast of saviors. The villains only act like villains, and the heroes only act like heroes. Never once are the heroes actions suspect, and never once are the villains actions sympathetic.

If you support the heroes of this article, and your dopamine lights up when your opinions are echoed in a publication, then you may love this article. Having said that, I am sure you have read this same article over and over again in many different forms, I certainly have.

Something I have noticed about tax and spend advocates is they have shifted their messaging from "these taxes will pay for great services" to "these taxes will hurt the rich". It's telling that even they have lost faith in the ability for tax increases to provide meaningfully better services as an advertisement. I suppose no one, including myself, would seriously believe it.

I am referencing only the American zeitgeist, I assume other countries might have better systems.

It feels more and more embarrassing as time goes by to tell people I am British.

Can you provide more information on this? I can't seem to find any.

What are the metrics this Australian law should hit? How do we know its achieving its intended result?

Countries like Singapore figured out that high reward / high risk (e.g. punishment) is the essence of capitalism and thus retaining talent in the public sector.

EDIT: You will notice a lot of talk about high pay. It's important to note that this is not without major punishment for financial mismanagement. You can't have one without the other, it's not just a question of giving politicians a bigger trough to put their snout in.

Yes, and to keep it like Singapore, lets increase the penalties for financial mismanagement.

Is there any evidence that this law is achieving the goals it was designed to tackle? If not, is there any reason it still exists? Why don't laws have to continually justify themselves as a matter of procedure?

A rather controversial opinion I have is this will begin to bend back very soon, once AI starts controlling more of the development process it will be able to reason about why we have these verbose frameworks at all, and start slimming them down to do only what it needs to do in the context of a program. I think stories like this will become more and more common: https://cybernews.com/security/curl-maintainer-stenberg-says...

Brexit is an interesting case to bring up. As someone who voted against it, it is also clear to me that it hasn't been implemented in any way that could have made it a success from the representatives, and looks like a clear case of the representatives maliciously complying with a decree that they didn't agree with.

I remain unconvinced that we can't do better than a system of government that is over 200 years old that only exists as a function of the technology at the time.

Keir Starmer is trying to sell this to public with the promise that it will curb illegal immigration. But by how much? And if that target isn't met, will it be rolled back? It always amazes me when governments hold themselves to such a low standard, and just fall back on "if you don't like it vote it out" while not recognising that the signal of a vote every 4 years can not meaningfully extract consent on this particular issue.

really interesting approach to calibration for hallucinations, im going to give this a go on some of my projects.

FWIW most of the inbound traffic to startups websites that I know enough to ask about is coming from ChatGPT.

This pops up every few years, and I bet once it gets in it never goes away. It seems asymmetric that one side only has to win once to win permanently while the other side has to win constantly. Is there any mechanism to stop this in the EU and make this kind of legislation explicitly barred?

I would hazard a guess that AI technology is already close to break even from an economic perspective, nearly everyone I know (especially those not in tech) uses it on a daily basis to assist with their work, especially coming in the form of dealing with pointless bureaucracy, writing emails and ideation.

From a cashflow perspective however it is no where near close to capturing its economic benefit, and to be fair to the bubble supporters I agree that I can't see how it can capture this in the short term. In the long term its pretty clear that there will be some major winners here who will reap big rewards.

Does anyone know what this fine is about? It says celebrity endorsements but I'm not actually sure what the root issue is here?

I don't know a lot about it, on the face of it I think its terrible. Why do you ask?

Diffusion of responsibility is definitely a defense in these cases, but the system should recognize this shortcoming and assign accountability (at least in an ideal world).

Although I'm willing to bet that the true actors here weren't necessarily presidents (even though they would ultimately be accountable like you say). Would be interesting to see who demanded what and when.

In an ideal world we would get to do a reverse investigation to understand which government officials were complicit in his very obviously politically motivated detention, action would be taken upon those individuals to ensure accountability, and the system itself would be updated so powerful interests can't abuse the law like this. How far are we from this world?

Is there any reason this couldn't be made into a right to prevent the constant legislative onslaught? Is it a lack of organisation/will or is it something intrinsic to the way the system is setup?

As an engineer you really have a finite amount of good working years, and accepting startup salary vs big tech compensation is a bigger risk than founders are willing to generally admit.

We should feel great that they now have a competent board that will make sure of that.

It is. Like I said above I would be concerned about the continuing success of the organisation according to its operating principles. Destroying the company over childish disputes would not be fulfilling that charter. I'm very glad OpenAI got the competent board they deserved after this event.

I'm not sure OpenAI compares to Enron or FTX in terms of unethical behaviour or financial fraud. The only similarities might be, funnily enough, disfunctional boards (or in FTX's case, no board).