“You have been deemed rich/important enough that we’ll allow you to bribe us”

Let’s be honest, that’s what this is. It’s rich people paying money to talk to those with political power. It’s a method of ensuring that those with political power can get a crapload of money in exchange for “listening to the grievances” of a select few. It’s an arrangement for connecting people with power and money more closely, while excluding the common masses from the rooms where decisions are made. It’s… corruption.

You’re absolutely right about that, and if we’re able to build a model that’s as capable as GPT and friends at parsing natural language, without simultaneously training it on everything from poetry to programming, that’s a major win. My current understanding of the field is that in order to build/train the models that are able to robustly parse natural language and “understand” the intent behind a series of instructions well enough to translate them to the correct tool calls, we need a very large and varied training set. I’m using “generalist” as a term to refer to the models that you can interact with in natural language across a wide variety of tasks. Those models are extremely powerful if you can also connect them to tools that solve problems deterministically, so that you get around the problem that they don’t really “understand” anything at all, while taking advantage of the fact that they’re extremely well suited for translating natural language to a selected set of pre-defined actions.

I think a major challenge going forward is that interpreting natural language requires a large set of training data. So training specialised models that can also interact with natural language is by nature difficult.

It shocks me that so many people are just blindly downvoting a comment like this. They make a very testable claim, and even cite a specific, easily searchable person, as their source. If you think the claim is unreasonable, it’s very easy to ask them for more background info or sources, preferably while pointing out why you’re sceptical.

To me, the claim that 16-20 year olds that are full to the brim of hormones, and have had less time to be exposed to various pollution than 30-somethings are more fertile seems reasonable off the bat. I have no doubt that I would have become a parent at 18 if it weren’t for contraceptives, yet I’m now closer to 30 and haven’t become a parent despite trying for a while. One side of that is that at 18 you’re more likely to be constantly pounding like bunnies, another side is that stress, pollutants, and probably a whole host of other factors make you less fertile with age.

My point is that we’ve built our model on top of these “generalist” models. You hook it up to an API and then let Claude, Mistral, etc. (I try to avoid GPT and some other) do the generalist job of translating human language into actionable tasks. You give it tools to parse documentation and actually do the tasks.

The generalist models are fairly good at taking a set of instructions and translating that to the correct tool calls, then our tools enforce correctness on the final output. Building an agent like the one we have would be nearly impossible without having some generalist model to do the “translation” step.

I think we’ll see two major changes going forward in how LLM’s are used: 1) they’ll become much more expensive and less widely used, since today they’re run at a loss. 2) they’ll be integrated into larger systems where they can do what they’re good at (parsing and outputting natural language), while offloading technical tasks to other tools that are actually built for technical tasks where formal correctness is paramount.

I don’t think they will. I’m the first to be massively sceptical of LLMs, but that doesn’t mean they can’t be used to build good tools. The key is recognising that tasks where correctness is vital should not be solved by the LLM directly. At my job, we’ve built an LLM-agent that’s very useful (internal use). What we’ve done is build essentially a Python library that this LLM uses to interact with our data. That way, we ensure that a query like “set up a skeleton for X” will be done correctly, while we save a bunch of time that would have otherwise been spent doing boilerplate work.

Basically: Enforce correctness by constraining how the LLM can interact with your data, and use the LLM to translate short natural-language queries into actions that otherwise would have taken 30 min of click-ops or write-run-toss scripting.

Show ads on the screen -> screen is vandalised -> nobody can buy gas anymore -> people go to the gas station without ads (and thereby functioning screens) -> lose money to the competition-> recognise that you lose money by blaring ads -> stop blaring ads -> … -> profit

This is why people like him hate/fear regulation. Regulating this is the people’s way (in a functioning democracy) to put hard barriers on what these people are permitted to do in order to squeeze out another dime.

aren’t bound to Evolutionary advantage to survive.

That’s not how evolution works though. Evolution is a process that works a the gene-level. A gene that makes its carrier more likely to reproduce and keep its offspring alive will over time propagate and replace genes that are less likely to do so. This is a simple game of statistics that works regardless of whether the organism that carries the gene is a human or not.

Basically, evolution isn’t about survival. It’s about what genes are more likely to propagate to the next generation. You can simulate this fairly easily: If you have a completely stable population (the average person produces one offspring), and a gene that makes e.g. 10 % of the population produce on average 0.99 offspring, you’ll see that after a certain number of generations that gene is drowned out and eventually extinguished. Any gene that isn’t extinguished has survived because it doesn’t put its carrier at a big enough disadvantage to be extinguished.

I’m not going to remember the exact domain of the survey company we use, what are you crazy?

I agree, and have decided to err on the side of caution, and also put the irritation over on higher-ups. If I get some link I’m required to click that I’m not actively expecting from an unrecognised address, just trash the email. A couple times, I’ve gotten follow-up from a superior asking me why I haven’t responded to <survey>, and I just tell them I haven’t seen it and that it probably got caught in my spam filter. They send me the link in question, and I respond.

I quite quickly realised that most of those surveys they need “everyone” to respond to will just slide quietly by when I do this, so I don’t need to spend time on them. My reasoning is that if it’s actually important, I’ll get it through a reliable channel, and so far that’s worked.

To be fair, I also dump anything that comes from some variant of “noreply” to junk. I figure that if I can’t reply, and I’m not actively expecting the email enough that I check my junk folder, it isn’t important.

In that case, ok I guess?

To be clear, your post reads like a legitimate question, especially considering the community you’ve posted to. I appreciate a good shitpost, but I tend to interpret questions on asklemmy and nostupidquestions as serious.

As a starter, I’m extremely sceptical of modifying the genome of embryos (as I hope was clear from my comment). Like you said, there’s plenty of defects that can have advantageous sides in addition to the negative sides. My point is that the only way this could even conceivably be ethical to do, is if it was used to remove diseases.

I believe there’s already a procedure they do to remove a disease carried in the mother’s mitochondria. I heard about a rare disease from my dentist the other day that causes people to never grow any teeth. I’m saying that if this is done, then the only ethical thing to do is use it for population-wide removal of severe diseases with no known positive sides, and even that can be dubious (as you point out).

I love ASOIAF, and honestly think GRRM is one of the (if not the) greatest fantasy writers of all time. He owes us nothing. He has created this fantastic universe, and a fantastic story. If anyone wants him to finish his life’s work, his epos, it’s him.

This kind of mentality that he somehow owes you to finish this story, and isn’t doing it because he “can’t be assed” is just so egocentric I don’t have words to describe it. Yes, I want him to finish it, probably just as much or more than you do. I also recognise that the man owes me nothing at all. He’s already given most of his life to this, and I’m enormously thankful for the universe, stories, and characters he’s already given us.

Sir

While I 100 % agree with the general sentiment that this is a terrible idea, I think your line of thought is a bit off. We have been made by evolution, a process built on the simple fact that any change that is too crappy will prevent itself from spreading, since the carriers of that change will be less likely to reproduce. Evolution is extremely efficient at preventing really crappy modifications from spreading. Thus, I don’t think our primary concern should be about these modifications “entering the common gene pool”. If they really are shitty modifications, their carriers will be less likely to survive/reproduce, and they’ll be watered down/wiped out by evolution.

I think what we should be worried about is twofold: First, it’s directly dystopic to imagine children born with defects that have been wilfully introduced, and which may not become apparent until in many years. Experimenting on unborn children this way is absolutely abhorrent. The second is the possibility that this actually works out in some ways, and is reserved for the super-rich, in which we can literally end up creating a human super-race that will inevitably suppress and exploit the rest of humanity.

Basically whether this works out or not, all outcomes look pretty bad. The only ethical way I could see this being done is using it to remove known defects such as hereditary diseases, and doing that through public programs aimed at eliminating those diseases at a population level. Doing something like that hinges on the program not being run by people like the one in the article.

“America!”

Based on how the question is phrased, it can be very open to interpretation.

I’ve thought all kinds of sick thoughts along the lines of “I could push this person in front of the arriving metro”, or “I’m holding a kitchen knife, if I were to stab the other person in the kitchen there’s nothing they could do about it”. When I was in the army, and we were at the shooting range, I could get thoughts about how easy it would be to shoot a whole load of people. Of course, I never consider doing those things. I’m talking about people I would lay down my life to protect, and I still get these thoughts.

I was very relieved when I read somewhere that this is fairly normal, and some researchers ascribed it to a protection mechanism: Their theory (which they provided some evidence for), was that be evoking these thoughts, your brain makes you more aware of acute potential dangers, so that you can act more carefully to avoid them. Basically, by becoming conscious that you could push/stab/shoot someone, you handle your movement/knife/firearm more carefully to avoid doing so by accident.

So: Have I thought about shooting someone? Plenty of times. Have I considered shooting someone? Never.

They don’t exist in my country, and to be frank, I’m shocked the ones in the US aren’t vandalised to hell and back by masked vigilantes at night.

Depending on how heavy simulations you want, it’s surprising how light hardware you can get away with in OpenFOAM. I used it for some university courses on a 2012 MacBook Pro (dual booted with Ubuntu) around 2021, and I could run 2D, two-phase simulations just fine.

Of course, if you want to run 3D stuff with large shear forces or turbulence and high time resolution you’re gonna have to grab a veeeery big coffee while you wait. However, the ability to stop and restart the simulation is really nice, and lets you see what’s been simulated so far.

Do it! It’s great fun to play around with!