dinfinity

I concur. Something to keep in mind is that it is often more robust to pull an LLM towards the right place than to push it away from the wrong place (or more specifically, the active parts of its latent space). Sidenote: also kind of true for humans.

That means that positively worded instructions ("do x") work better than negative ones ("don't do y"). The more concepts that you don't want it to use / consider show up in the context, the more they do still tend to pull the response towards them even with explicit negation/'avoid' instructions.

I think this is why clearing all the crap from the context save for perhaps a summarizing negative instruction does help a lot.

The solution is almost trivial and well known:

Document. Whenever a member of the team leaves, you should always make them document what is in their head and not already documented (or even better, make documenting a required part of normal development).

The thing to understand is that every (new) AI chat session is effectively a member entering and leaving your team. So to make that work well you need great onboarding (provide documentation) and great offboarding (require documentation).

> autophagy [...] usually takes at least 18 hours of continued fasting to even start

False. That would mean cellular debris would pile up unconstrained for pretty much everybody, which is clearly absurd.

> It is also an extreme cell response that is associated with high levels of cellular stress

Also false. It is a very essential cellular process. Read up on it, please.

The 'comeback' of racism? Really?

It never left. Anywhere in the world.

Ignore the insanity of banning books for a second. What about the Victorian-era insanity of preventing children from being exposed to the mere concept of sex and labeling it 'obscenity'?

On one hand we want kids to learn about consent, what 'normal sex' is like and all that, but simultaneously there is this idiotic push to prevent them from encountering any of it until they are 18. If we don't want kids to see bad porn, we need to ensure that there is lots of good porn available, and not just some boring sex ed bullshit. I mean actual benign everyday sex that kids can safely watch and learn from because otherwise they will never see it anywhere else (it's not like they regularly watch their parents or other people do it).

You have to be incredibly regressive to think 18 is somehow a good cutoff for this.

I would say:

1. Spend time on the fundamentally new things through whatever tool provides it.

2. Always try to prevent 'vendor' lock-in. Think about portability and reusability.

A lot of the stuff that now "works well" when working with AI assistants and tools is stuff that was always a good idea and always worked well. Write Once, Read Many; Provide good specification; Communicate clearly; Automate repeated tasks; etc.

If you update your workflow in a fundamental manner and don't jump from investing 100% in tool X to 100% in tool Y redoing a lot of shit, you improve it efficiently.

LLMs don't learn, though.

The fundamental problem with social media (and many other things) is humans, specifically our biological makeup and (lack of) overriding mechanisms. One could argue that pretty much everything we call 'civilised behavior' is an instance of applying a cultural override for a biological drive. Without it, we are very close to shit-flinging murderous apes.

For so many of our problems what goes wrong is that we fail to stop our biological drive from taking the wheel to the point where we consciously observe ourselves doing things we rationally / culturally know we should not be doing.

Now the production side of media/content/goods evolves very fast and does not have a similarly strong legacy biological drive holding it back, so it is very, very good (and ever improving) at exploiting the sitting duck that is our biological makeup (food engineering, game engineering etc. are very similar to social media engineering in this regard).

The only reliable defense against that is training ourselves to not give in to our biological drives when they are counterproductive. For some that might be 'disconnect completely' (i.e. take away the temptations altogether), but having a healthy approach to encountering the temptations is far more robust. I am of the opinion that labeling the social media purveyors and producers in general as evil abusers is not necessarily inaccurate, but counterproductive in that it tends to absolve individuals of their responsibility in the matter. Imagine telling a heroin addict: "you can't help it, it's those evil dealers that are keeping you hooked to the heroin".

We could also point the finger towards the popular consensus that "salt is bad for you". This guy just took it to the next level.

Instruct it to look at the actual code of your dependencies directly. It should be able to (quickly) figure out how the exact version of the dependency you have should be used.

Just like the term "ultraprocessed foods" the term "sandwich" is very badly defined.

I can't determine whether the study differentiates between healthy home made sandwiches and junk food that contains some kind of "bread".

Human societies and technology are, to our current knowledge the most advanced, ordered, low-entropy and complex things in the entire universe.

Whether the universe as a whole evolves (pockets) in that direction as some fundamental goal is a philosophical question, but that it does so is clear.

Let's remember that the basic defense against model collapse is just not training on AI-generated and other crap data.

Sure, there are places where determining whether it is AI-generated or 'real' is hard, but there are plenty of places where the trust in the provider provides enough basis to include the data during curation. For example, it's not as if the NYT will suddenly start pumping out unchecked AI slop.

And then there is the enormous potential of data synthesized aided by, but not completely generated by AI and validated for accuracy through systematic means.

I'm not sure tit-for-tat works with organisations. If leadership changes, the trust that you built up can just go out the door with the old leadership. The EU is learning that the hard way.

What remains stable is a very basic and primitive power dynamic (generally in resources and violence).

I don't think the "connecting the world" was the problem. IRC also has tons of toxicity and connects people all over the world.

The core problem is gamification of social interaction. The 'Like' button and everything like it for things people say or show is hands down the worst thing to happen on the internet. Everywhere they can, people whore for karma (unless they spend a lot of mental effort to fight back that urge). How primitive the related moderation systems are directly affects how much primitive shit gets rewarded and alas, most moderation systems are ridiculously primitive.

So, dopamine hits for saying primitive shit.

False. If the Commission accepts it, there is still time and room to fight it. Furthermore, your logic would also apply if some random person on the street proposed this.

My point is that legislation can exist in a variety of stages. In the extreme case where we get stressed and angry at every proposal that any politician comes up with, we waste our energy instead of letting the specific systems introduced and thought up to mitigate issues with such proposals do the job efficiently.

Said otherwise: In some cases it effectively becomes crying wolf and makes people numb to and averse to hearing about it. As said, though: I am ambivalent towards news like this. In some cases we need to 'tell, tell, tell' to get the point through. It's a fine line, I think.

I am ambivalent about news like this.

On one hand it is a terrible proposal that deserves exposure to generate pushback.

On the other hand it is still just a proposal, of which there are many (stupid ones) which never see the light of day. This still needs to pass the European Parliament (which is generally opposed to this, not in the least because it is very unpopular under EU citizens). Even then the chances of it surviving a challenge in the courts (of which there will be many) are tiny.

> Ok but what happens when unhealthy kisses healthy? Does it balance out?

If you're lucky, sparks fly ;-)

> If it's not risky, I'm about to embark on a mission.

"Hey baby, can I sample your oral microbiome?"

For people who don't like drinking beetroot juice, you can also regularly (French) kiss with people with healthy oral microbiomes [0][1]. Best to really get in there.

[0] https://www.adelaide.edu.au/newsroom/news/list/2024/02/20/mo...

[1] https://microbiomejournal.biomedcentral.com/articles/10.1186...

I'd like to add: Stretching or some other kind of flexibility improving activities. Muscles moving for half an hour a day doesn't (necessarily) do anything for that (may even make it worse if you're doing heavy stuff).

The effects on quality of life of a bit of flexibility are huge. Back pain, knee pain, shoulder pain, "RSI", and so many other ails are often just pretty much permanently cramped muscles negatively affecting ligaments and nerves.

> Is this a problem with React itself?

React is far from the best tool for most jobs (if any), but it has a shitload of inertia.

I can highly recommend looking into Lit (used to be Polymer): https://lit.dev/docs/

IMHO it allows just the right amount of encapsulation and structure, whilst building on W3C web components and being highly interoperable with other javascript libraries/vanilla code. It's pretty much what web components should be out of the box.

This answers most of the important questions surrounding this specific tech. Much appreciated.

> The WHO did not word it badly, it was the accepted knowledge back then.

Bullshit. I already showed how they were careful in the rest of their wording in your own source. 100% efficacy was never 'accepted knowledge'.

> Why is that on me? You tried a stab at antivax whackos and it backfired

Bullshit. The base rate fallacy was and is very present. The misinterpretation I referred to made sense to mention: It is a very commonly known recent and relatable example of the base rate fallacy. You made it a tribalistic ad hominem in your own head.

> you casually remark that its the low efficacy of the vaccine, not being unvaxxed, that causes the majority of people who end up hospitalized.

I did not say this at all and you still do not understand the base rate fallacy. A larger percentage of the unvaccinated people were hospitalized than of the vaccinated. There were just generally way more vaccinated people.

> Maybe we should have worked on a vaccine with better efficacy (hey, let me dream up the impossible, maybe immunity like with the measles vax where 99% ppl actually can't get sick at all) instead of harassing unvaccinated people?

See, now there we go. The mention of the Covid-19 vaccination triggered this frustration and made you decide to inject it where it did not belong and keep your mind closed to the math and science surrounding this. The math does not care about your feelings and is what it is. Accept it.

> If i quote the WHO[1] on this, you might just say they are not literate enough.

This is a page primarily about herd immunity, not about vaccine efficacy. You had to dig to find this (or somebody dug for you).

Having said that, even though they added several sentences on how certain things are unsure or need more research, I will admit that they worded this specific sentence badly: "Vaccinated people are protected from getting the disease in question and passing on the pathogen, breaking any chains of transmission."

It is clearly overstated and does not match the careful wording later on: "We are still learning about immunity to COVID-19. Most people who are infected with COVID-19 develop an immune response within the first few weeks, but we don’t know how strong or lasting that immune response is, or how it differs for different people. There have also been reports of people infected with COVID-19 for a second time."

> The rest of your post are personal attacks that do not add to your point.

They do, because you clearly have a bone to pick that is preventing you from rationally approaching the matter and discussions in general. Think about what point you're actually trying to make and what that has to do with the base rate fallacy. Really, verbalize it. What is it? Why did you feel the need to inject that into the discussion, even though it doesn't belong here at all?

You must be a liar or willfully ignorant to say that after the entirety of 2019-2021 happened. The efficacy of vaccines in general and specifically of the various Covid-19 vaccines have been talked about ad nauseam. No even merely scientifically literate person has said that the Covid-19 vaccines (or any vaccine for that matter) are 100% effective.

It's so weird how people will close their eyes for basic science to virtue signal to their group. I sincerely hope you open your mind and prevent your virtue signaling from killing you (or anyone you know) in the next pandemic.

> In and of itself own, as quoted by OP, it’s not meaningful.

This is an overly literal, pedantic, and ungenerous interpretation of what I said.

Clearly the actual numbers as well as the base rates for the relevant groups are necessary to meaningfully interpret the statistic, but I was hinting at a similar case after describing the logic for the original case, not exhaustively describing the similar case.

Cool, so your question was in bad faith, you were not at all prepared to learn, have wasted our time, and you still don't understand anything about the base rate fallacy.

Honestly, what a shit comment.

I personally would like to be able to choose specifically for a male driver. I trust them more.

I'm being facetious of course, but it does highlight how bad sex discrimination is.

The problem with sex discrimination here is that it is a very crude workaround for a fundamental problem, namely not being safe with some Uber drivers. I'd like to point out at this point that as a dude (especially a scrawny one) you can also get murdered or (sexually) harassed by an Uber driver (male or female). This workaround does nothing for those cases. The fundamental problem will probably simply be properly solved with driverless taxis.

> I think The Expanse did a much better job of modelling the reality of future economics than trek ever got close to.

That is because The Expanse does a lot of "the stuff that happen(s)(ed) on Earth, but in space!". Don't get me wrong, it also does a lot of great scifi stuff, but the factions and people are quite one-dimensional unimaginative analogues of known factions.

This approach makes it relatable (and commercially more successful) but not necessarily more realistic. It's like predicting flying horse carriages and flying cars versus helicopters, planes, and rockets.

Related: IMHO, one of the worst things about the 'relatable extrapolation of the present' aspect is that it limits popular scifi enormously. There's usually some special space carved out for humans or very human-like creatures doing very human things with the environment pretty magically being incredibly Earth-like all the time for hundreds or thousands of years in the future, even though the lives of humans today are already incredibly alien compared to those of humans just 200 years ago.

Exactly. Broken ideologies and populism break down very fast in the face of proper debate, but that debate needs to be had and broadcasted very very loudly and broadly; Otherwise it is drowned out by the unchallenged propaganda.

No, this is incorrect and misleading. It is a meaningful statistic when interpreted right. See sibling comment.