The limiting factor of development (for money) is good ideas, or valuable ideas. If you have 5 features you want, but only time in a month for 2, then that forces you to choose the best 2, which come with the tech debt, support and opportunity cost.
Removing the opportunity cost doesn't eliminate the other two costs of a feature
A few years back if you asked me to build a system and just name the price, my question would be: What should it be able to answer? Whatever i want. I wouldn't have an idea how to build a system like this.
Today we do.
Would you add any context to your comment to even have an argument we can discuss?
If Astra is AGI, then GPT 4 was AGI too. There's no qualitative difference. It can write poetry and role play with you, and code snippets, and you can run those code snippets if they're in bash format and boom, you have an agent.
But if that's AGI then it's a meaningless term; it's just a martketing gimmick. Which is entirely my point.
There is a significant qualitative difference, very obvouse one too.
The complexity of tasks GPT-6 can do are not compareable to GPT-4. We see this daily. From CVEs which became a real problem after GPT-4, generic tool use (blender, paint, etc.), agentic work (GPT-4 was at a minute or so, GPT-6 can do hours).
They broke through significant barriers in a short period of time. I would def count the agentic quality of doing hours of work as one of the main proofs.
I can tolerate Teams being a bit mid because the core functionality nearly always works. I can tolerate Outlook being meh because it's doing a lot (nothing I, a lowly IC really need though I can imagine the top hatted boss who does) and I think it does okay with the scope it's tackling.
But the M365 copilot app is inexcusable for being as core to Microslop's strategy.
It's such a broken POS. Code snippets, especially for html, frequently don't render, or if they do, they render as rendered html which will frequently be an empty box, and you have to manually (for each snippet box) switch to show code mode.
Jevon's paradox isn't a physical law, it doesn't magically apply to everything. Millions more copies of Atari's ET game didn't cause everyone to pickup a cheap copy, and cause extra demand for a garbage video game. Some times (actually, usually, I'd argue) things are made that will sell for less than the cost of construction because of irrationality, and they don't induce extra demand and they don't change the negative profit margins.
You can't simply wave Jevon's paradox at things. Thousands of miles of canals were dug in the UK that couldn't be sustained and were abandoned. Thousands of miles of railways were laid that could be sustained and were abandoned. And those are potentially durable investments, unlike cheap walls, pillars and roofs laid over a levelled concrete slab full of fast depreciating IT equipment.
I'm so glad the tide here is turning on this talking point, brought on by exactly the same people beating us over the head with it for months while no progress is made towards it materializing.
Many, many people who post here are capable neither of real analysis nor distinguishing real analysis from memes. They aren't hackers, they are adherents of a cult that happens to focus on the same subject matter as hackers.
Jevon's paradox only applies to products with near infinite demand. Energy being the most famous example. I dont think its difficult to argue that compute/intelligence is also a base input into the economy and theres almost no limit to the amount of intelligence the world will want.
The article observes that the cost of frontier intelligence from 2025 has fallen 100x in the last year. It also notes that the energy to run models is also collapsing. Consumer hardware is borked right now because these new algorithms are revolutionizing the utility of a computer. Computing is technology who's cost has been collapsing for 90 years, and its a safe prediction that it will decrease again.
> Yeah okay bud, anyone checked in with the state of consumer hardware recently? Not the author, evidently.
RAM prices will crash when demand drops even a little. They'll probably crash to a lower (inflation adjusted) level than before. This has happened before.
Industrial scaling in general often looks like a sawtooth: price spike, capacity investment, crash, repeat.
Part of what's keeping prices high a little longer is that everyone knows this and is a little reluctant to plow resources into chip fabs for fear of having the bottom fall out before they recoup or sell that to someone else to hold that bag.
Graph the average compute and RAM in a mid-high end laptop at an inflation adjusted price point for the past 40 years. It's very exponential and hasn't slowed down much.
No. Prices will crash when supply side expands to meet the increased demand. Because demand won't go down to pre-bubble times any time soon. Unfortunately the supply side has been very slow in increasing production, partly because most steps of the production chain are all maxed out.
On a long enough scale you are right that prices will likely normalize to a better level, but before 2030? That would mean the factories are built quickly once they begin.
The entire reason why hardware prices are so absurd right now is because manufacturers across the board are doing everything to prevent that crash.
They all collectively chose NOT to increase supply with increased demand. So if the bubble pops, they just go back to previous prices without oversupply driving the prices to rock bottom.
There's been multiple YouTubers saying they've been paid recently to make... AI Doomerism videos (e.g. Sabine).
I imagine this is simply part of the campaign, AI doomerism followed by AI Bloomerism, it could even be part of the same organisation's campaign; selling the solution to a problem they put in your mind. Hype, criti-hype, it's all so tiresome.
AI doomerism more importantly I think legitimises current AI as being super powerful and useful. If something can take over the world then surely it's a powerful technology !
>could even be part of the same organisation's campaign
Eh we know its part of the same campaigns. "Please provide us with regulatory capture because our tech is an existential threat" has been going on for a while now.
> people writing in a non native language
This category of people are going to be most betrayed by LLM output: 1) the receiver loses the signal that the sender might need to be queried to find their real intent and 2) the sender doesn't have the ability to determine if what they are sending is what they mean
> people insecure in their writing
These people can grow up, I don't care. Not a good enough reason to send a slop grenade.
>people not used to writing in industry terms
Similar to the non-native speakers, but slightly less in magnitude. They can educate themselves though.
- people with the curse of knowledge that are aware that they can’t write for a general audience well
These people probably can get some value out of it but they should take care
Btw calorically dense foods are, in fact, hyper nutritious. While you may imagine healthy foods to be green vegetables and raw meats, you can buy - from regular supermarkets - roast chickens ready to eat, salad bowls and fresh fruit. You need not even store it, but buy and eat on the go.
People repeat the myth that healthy food choices are expensive (they aren't) and time & space consuming to make (they can be) so often that most people believe it. But it just isn't so: you can eat at McDonalds every day, and if it's only a part of your diet, it's healthy. If you can get fibre and some vitamins elsewhere, and aren't overeating, the protein and calories in a fries and a hamburger are fine. Good, actually.
Healthy foods are simply principally marketed to, sold at a large markup to, middle to upper class Americans so they believe that's the usual market value.
Now, it can be expensive in time, effort and equipment for struggling, hard working, irregular shift workers. That's a different issue though.
Removing the opportunity cost doesn't eliminate the other two costs of a feature
reply