Swiss

@MrFroggy42 @RonSwanonson Expensive compared to what?

@RonSwanonson Too slow, too expensive. I really don't get the bitcoin hype. Having studied this stuff quite a bit I just don't see how it lasts. It has/had first mover advantage sure but once you start talking about tokenized assets on a ledger bitcoin starts looking useless to me.

Had a hotel bar convo with normies last night. They were all history buffs, and they asked if I liked history… I told them I’m not a huge history buff but that I knew a lot about the history of money, the different types of money and how ultimately, central powers were always able to capture the money at the expense of the actual value producers in every society we’ve had etc. I pointed to the fact that money is a language that can direct the course of history more than any other factor They were all blown away that they knew so much about history but had never really thought about monetary history and how it constantly changed the course of humanity One person even said “oh are you into cryptocurrency and Bitcoin?” And I said, well yes this all does lead to Bitcoin as the first money that can’t be captured but I’m not going there, and stuck to monetary history. With their reactions, l guarantee that those 4 people left with an itch to learn more about the history of money Once you understand the history of money, you won’t have to pitch “Bitcoin” I’d even suggest to stay away from the word altogether and plant curiosity that ultimately leads to Bitcoin anyway

@Choosey_Beggar @RonSwanonson Mine isn’t capturable.

@RonSwanonson Is bitcoin not being captured as we speak?

@0xKofi As of now, you need to review it. But as the tools get better about asking the right questions, the need will go down. It’s currently so frustrating how badly they assume things. Just ask me, Claude!

I've been seeing a lot of folk on the timeline say IDEs are dying. But I don't get it. Are you guys not reviewing the code your agents generate 😅

@DeryaTR_ Not for the plebs, though.

In the next 2–3 decades, we’ll achieve more technological progress & economic growth than in the entire history of human civilization! Since the dawn of civilization, we’ve created ~$6 quadrillion in wealth cumulatively. By 2050s, we’ll generate another $6 quadrillion in wealth!

@squigglydonut @ChrisJBakke Done? He has 3 more years, sadly.

@ChrisJBakke Trump is done Americans don’t support him only old boomers

Japanese reporter: "We are allies. Why didn't you tell Japan about the Iran strikes?" Trump: "Who knows better about surprises than Japan? Why didn't you tell me about Pearl Harbor?"

@SnerdlyP @scottsantens So in the future, when robots do all jobs, we should make humans scrub things?

@scottsantens If we have UBI, then I want UWR (Universal Work Requirement). If you aren't employed, then you show up to clean highways, clean parks, scrub graffiti, etc. No work, no food.

I don't like talking about the cost of universal basic income, not because it's some high number, but because it assumes not doing it is FREE. We spend $1.5 TRILLION a year on the costs of child poverty alone. We cut child poverty by 40% for a cost of $100B in 2021. That would have been a $500B GAIN, but Congress decided that saving that much money cost too much. It's as dumb as saying that buying $100 for $20 is too expensive. What are we doing? We are 💯% fine spending any amount of money on wars and tax cuts for the rich, but I'm the one who has to defend spending public money in a way that reduces overall costs? I shouldn't have to convince anyone that investing $100 to get far more than $100 in return is a good deal. There are so many ways of going about UBI, and thus so many different price tags. It's not a simple answer. But what is simple is understanding that poverty, mass chronic insecurity, and extreme inequality all impose huge costs on our society. The cost of UBI is lower than the cost of not doing it.

@Shyguymask99 I loved paint.net. Unfortunately, I gave up on Windows years ago.

I feel validated that people hate GIMP now. Like 10 years ago in the 2010s everybody would recommend it like it's the best software ever, but when I tried it I found the UI insanely shit & confusing to work with. I quickly went back to paint .net and I still use it to this day.

@thinkingshivers Obviously it will improve rapidly, as it has been doing.

Not enough people talk about how unpleasant vibecoding is. The best analogy I can think of is driving. It's cool that we can just hop in a car and drive to the store. It's a lot faster than walking. And yet, it's so stressful and infuriating, we had to invent a new word just to describe its effect on people: "road rage." AI-assisted coding is the same. It's so much faster--there's no going back to coding everything by hand, the equivalent of walking everywhere. And yet it's incredibly annoying and stressful. It's characterized by annoying delays between requests, time-wasting misunderstandings, blatant lying, and absurd overconfidence. Hopefully this gets better as models improve.

@agmethod75 @tomfgoodwin Or contexts.

@tomfgoodwin Theyd have been better off calling them threads or processes.

I’m surely being stupid. But if AI is rather unconstrained by expertise or capacity or to some extent speed Why do we need to divide tasks or departments to 9 agents ( the marketing agent, the optimization agent etc ) to each do one thing. And then another agent to manage the swarm. Cant one agent just be doing it all you know. It seems very skeuomorphic. Will we have HR agents to make sure the agent agents are being looked after ? A office canteen manager agent to feed the agents ? Seems daft

@tim_tyler @tszzl Doesn’t that depend on the teacher AI being aligned?

@CremlinGremlin @tszzl Well, there's also RLAIF: anthropic.com/research/const…

modern alignment methods seem to work reasonably well across orders of magnitude of model scaling, survived the transition to verifiable rewards and that should at least inform your decision making

@tszzl @tim_tyler The post-training is mostly what, then?

@tim_tyler we do not solely, or even mostly, rely on human feedback

@Ricburton @tszzl Respectfully, what is your point?

@tszzl Have you ever camped in the desert or hunted for survival?

the romantic human aspect of space colonization-placing our flags on other worlds-is redundant with robots and Von Neumann replicators unfortunately. the mars colony will be robots. ppl can join, but as consumers of an experience rather than critical parts of novel adventures

@svpino Plus, the AI gonna keep getting better.

Let's be honest, most people ain't checking the code they're writing with AI. Yeah, it may be important, and we are always yapping about how you gotta do it, but there's absolutely zero chance most people are taking the time to slow down and go line by line, trying to understand everything the model wrote. It's just too easy to produce a billion lines of code now. It's even easier to test the product (rather than the code) and ask the model to fix whatever you don't like. And you know what? I think not looking at the code is fine for many. I've seen a lot of code 100% written by humans. Some of it is an immense pile of garbage, and nobody has died. 100% AI-generated is actually an improvement for those products. But there are many places that can't afford slop, and we always take things too far. Non-supervised AI-generated code is dangerous. There will be some blowback, and companies will start getting very wary of AI cowboys. Some might outright ban AI-generated code, and some will figure out how to make developers liable for their code.

@liron Why not buy a second $200 account? They can work on the same repo.

Claude Code Max plan costs $200/month, and is a great (subsidized) value. There's also a feature where you can spend approximately $100/hour to make it go 2.5x faster. I am using that feature.

@Yuchenj_UW Well, probably not xAI.

Some people at frontier AI labs told me they believe startups are over. OpenAI, Anthropic, Google, xAI will absorb every industry as AGI nears. Coding today, science, medicine, and finance next. Then everything else. If they’re right, that’s a pretty boring end of the world.

@snaxion @BitsagaRob No. Either you weren’t there or you didn’t the research. For me, it was the latter.

@BitsagaRob Back then we didn't know if: a) it would work at all, long term b) it was illegal

This man was mass-producing millionaires in 2013 and nobody showed up

@molecularmusing I loved writing code (for 40 years), and I was good at it. But I now see that we have entered a new era. Those days are over.

I find this extremely worrying, with many of people I respect saying things like "I no longer write code" or "let LLMs do it". Why did you start programming? Was it never the journey for you, but only the goal? I genuinely want to understand this, I seem to be the odd one out.

@sama Why do you always sound so douchey?

I have so much gratitude to people who wrote extremely complex software character-by-character. It already feels difficult to remember how much effort it really took. Thank you for getting us to this point.

@MarkovMagnifico It is pretty easy to get to a near-optimal solution. The hard part is proving that your solution is 100% optimal, which is not necessary in the real world.

it's honestly crazy the number of salespeople I've talked to who have a literal traveling salesman problem in their work and they're always saying, "There's no good software for this!"

@BobMcElrath @ProjectJupyter Dependencies? Never touch the stuff.

The package dependency hell is a lot worse with AI. I just failed to install @ProjectJupyter with pip due to a half dozen broken dependencies. I used AI to fix it. So is everyone else. Maintainers pin dependencies and have no incentive to make APIs or backwards compatibility.