richardw

That applies to all computers when you buy the fully specced versions on day 1. A maxxed out iPad isn’t throwaway, but framework represents a higher standard of upgradability.

Isn’t this an issue with eg Cloudflare removing a portion of the web? I’m all for it from the perspective of people not having their content repackaged by an LLM, but it means that web search can’t check all sources.

Got a killer original idea, got it built, learned a ton, is embarrassed about a few failures because he’s actually a solid professional. Well done, mate. If it didn’t cost you $100k, take down the network or lose you your job, you’re ahead and will laugh in a year.

If the system knew both sides were the same vendor or used the same algorithm, would it be better to stream the scene/instructions rather than the video?

I suppose the issue would be media. Faster to load locally than push it out. Could be semi solved with typical web caching approaches.

Never write your own multi-planetary date/time library in brainfuck.

Scary part is: what if it’s inevitable? We don’t get to choose our environment, and toss one is forming around us.

A friend’s dad only knows assembly. He’s the ceo of his company and they do hardware, and he’s close to retirement now, but he finds this newfangled C and C++ stuff a little too abstract. He sadly needs to trust “these people” but really he prefers being on the metal.

I got a pair of Pro 2’s in 2022, still going amazingly strong and I hammer them. I often fall asleep with them in.

Only weird thing is that they sometimes need to be put into the charger and taken out again to for one to connect. It just seems they should be able to reset without the ceremony.

https://freedom.to/ Works everywhere for me.

Converted into energy and then escape the black hole, from which light can’t escape? That doesn’t seem to compute. And if it’s converted into gravity waves then we have an excellent obvious candidate for how most energy will escape a black hole. It won’t be waiting around for hawking radiation.

Testing the waters. I’d guess if it pays any dividends things will ramp up. This is also across all suppliers. At some point they’ll put more wood behind fewer arrows.

Housing costs, cost of education for yourself and kids, needing to work and study longer to provide more. More women want careers longer.

Basically we compete more until everything costs more. Now we need multiple salaries because everyone we’re competing against is a highly paid couple. Rinse and repeat so everything of value is too expensive for “average people”.

Countries are placing their bets. Fossil fuels will be a massive waste of investment in a decade. Anyone who can extrapolate a graph sees where this is headed.

Right now you’re putting in unrequested effort to get to an answer. Nobody is driving you to do this, you’re motivated to get the answer. At some point you’ll be satisfied, or you might give up because you have other things you want to do, more.

An LLM is primarily trying to generate content. It’ll throw the best tokens in there but it won’t lose any sleep if they’re suboptimal. It just doesn’t seek. It won’t come back an hour later and say “you know, I was thinking…”

I had one frustrating conversation with ChatGPT where I kept asking it to remove a tie from a picture it generated. It kept saying “done, here’s the picture without the tie”, but the tie was still there. Repeatedly. Or it’ll generate a reference or number that is untrue but looks approximately correct. If you did that you’d be absolutely mortified and you’d never do it again. You’d feel shame and a deep desire to be seen as someone who does it properly. It doesn’t have any such drive. Zero fucks given, training finished months ago.

They’re great at working with the lens on our reality that is our text output. They are not truth seekers, which is necessarily fundamental to every life form from worms to whales. If we get things wrong, we die. If they get them wrong, they earn 1000 generated tokens.

I’ve wondered whether humans who wanted to protect some areas of knowledge just start writing BS here and there. Organised and large scale, with hidden orchestration channels, it could potentially really screw with models. Put the signal to humans in related but slightly removed places.

Nah I’m all over the place. I said the last 10, to check if the 90% claim could be true if you do what I’ve done recently: use it for tons of little general ad hoc things rather than eg code needing serious accuracy.

I don’t “trust” it in the way I’d trust a smart colleague. We know how this works: use it for info that has a lot of results, or ask it to ground itself if it’s eg new info and you can’t rely on training memory. Asking it about esoteric languages or algo’s or numbers will just make you sad. It will generate 1000 confident tokens. But if you told me to lose Google or ChatGPT+Claude, Google is getting dumped instantly.

It ranged from whether an epic v10 sport surf ski was a good fit for a newbie, to entra ID questions, to local data residency compliance laws, new jira alternatives, why schools ask for closed shoes, text to speech tool search. Many of these I use eg o4-mini-high for because I want it to ground itself: find material and compile something for me, but get me an answer fairly quickly.

Ones that don’t work but weren’t in the last 10: voice. It sounds amazing but is dumb as rocks. Feels like most of the GPU compute is for the media, not smarts. A question about floating solar heaters for pools. It fed me scam material. A question about negotiating software pricing. Just useless, parroted my points back at me.

I scale models up and down based on need. Very simple: gpt-40. Smarts: o4-mini-high. Research: deep research. I love Claude but at some point it kept running out of capacity so I’d move elsewhere. Although nothing beats it for artefacts. MS Copilot if I want a quick answer to something MS oriented. It’s terrible but it’s easy to access.

Coding is generally Windsurf but honestly that’s been rare for the last month. Been too busy doing other things.

I just went through the last 10 chat titles and all of them were spot on for me. Maybe the person you’re responding to has a different experience than you do and calling their perspective “suspect” is somewhat uncharitable.

(There are times I do other kinds of work and it fails terribly. My main point stands.)

Surely you can’t be too surprised. The market is pushing for move-fast high polish, speed over substance. You can just do things, move fast and break things, etc. Velocity is the moat, indeed.

This is the market YC is breeding. When these guys float to the surface, what did you think would happen?

YC, you’re one of the greatest generators of value ever. Do better.

Interesting. I’m only a user of them but not had one second of trouble. I save them on my device in the native saving place (iOS/mac) and it just works. I didn’t know this issue existed and I’d like to avoid it. Is the issue when you save them in a password manager?

I have Bitwarden for personal and now 1Password for work, so might hit the issue at some point.

I’d considered it but assumed a person in one place is cheaper than many people wandering around to fill their own cars. Just bake it into the price. This is not an insurmountable problem compared to “require all buildings must be upgraded to power EV charging”. If I paid you a million dollars to figure it out I think you could make a plan in under a day.

The benefit is that you can power the cars when there is least demand on the grid, on the roads, on the need for the vehicles themselves, on the time of people who currently wait at chargers for their cars to charge. Fix one thing (and build upon the main innovation of cars driving themselves) and you unlock all this other waste. The charging location can even manage this by setting availability to align with staff, eg 5am to 11pm, or 24/7 if they want to hire up. You can even have roaming staff, eg wander between three locations and just switch the cars out.

I’m making very loose claims. Hope, it’s not an huge stretch, etc. My error bars are very wide, intentionally, because it’s hard to predict the next 5 years. Your position is a lot more brittle. If self driving works, cars can go fill themselves up. That can change demand, so the demand argument falls away. If taxi’s work and scale up, costs drop, so the economic issue falls away. For you to be right, no innovation must happen.

The point I’m primarily trying to make, repeatedly: distance to a charger is not some universal rule that prevents uptake. The least likely thing to be true is that the market stays the same. If it remains the same in the US for a few years, China will crush this market.

My comment is not solving for people who would be better served by charging at home, because they don’t care about the distance to a charger. Waymo’s can already find their own way to a charger, so it’s not a huge stretch to imagine people innovating around this problem for owned cars.

It’s not insane to hope many cars will be able to go and charge themselves at 2am, when roads are quiet and chargers are free, a few years from now. Optimise over the entire system, schedule it, car ready for the morning.

Not only the climate situation, the economic situation. If the US protects the old tech for another decade it’ll never catch up. The US needs to move along the experience curve as fast as possible, build skills and volume and charging stations and suitable power grids and sources. I would much, much rather be China than the US in this fight right now.

“We went outside this weekend. Terrible. I wasn’t hot anymore, the smog was everywhere. House was tiny. No AI to help with conversations and people were unfriendly. I’m staying plugged in, where we can fly amongst the stars on unicorns. Some say it’s fake but I say life has been fake for a while.”

Survivorship bias. The guys going home at 5 went home at 5 and their companies are not written about. It’s dark but we’ve been competing for a while as life forms and this is “dark-lite” compared to what our previous generations had to do.

Some people are competing, and need to make things happen that can’t be done when you check out at 5. Or more generally: the behaviour that achieves the best outcome for a given time and place, is what succeeds and forms the legends of those companies.

If you choose one path, know your competitors are testing the other paths. You succeed or fail partly based on what your most extreme competitors are willing to do, sometimes with some filters for legality and morality. (I.e. not universally true for all countries or times.)

Edit: I currently go home at 5, but have also been the person who actually won the has-no-life award. It’s a continuum, and is context specific. Both are right and sometimes one is necessary.

If I want facts that I would expect the top 10 Google results to have, I turn search on. If I want a broader view of a well known area, I turn it off. Sometimes I do both and compare. I don’t rely on model training memory for facts that the internet wouldn’t have a lot of material for.

40 for quick. 40 plus search for facts. O4-mini high plus search for “mini deep research”, where it’ll hit more pages, structure and summarise.

And I still check the facts and sources to be honest. But it’s not valueless. I’ve searched an area for a year and then had deep research find things I haven’t.

Photoshop etc are still just tools. They can’t beat us at what has always set us apart: thinking. LLM’s are the closest, and while they’re not close they’re directionally correct. They’re general purpose, not like chess engines. And they improve. It’s hard to predict a year out, never mind ten.

CEO’s will be fine until their customers disappear. Are the AI’s going to click ads and buy iPhones?