Indeed. I for one enjoyed this piece. Yes, it had errors and lots of odd grammatical choices, but the reading remained affordably challenging and the prose had a newness to it.
> Seems like I am engaged in the ice trade and they are about to invent the refrigerator.
The way I like to look at it is that I'm engaged in the ice trade and they are about to invent everything else that will end mine and every other current trade. Which leaves me with two practical options: a) deep despair. b) to become a Jacks of all trades, master of none, but oftentimes better than a master of one. The Jacks can, for now, capitalize in the thing that the Machines currently lack, which is agency.
When it comes to cancer, there is an awful lot of legacy thinking and "way things are done" taking lives. Starting with the so called "standard of care", which makes patient lose precious treatment windows while they wait for a possible miracle from "first-line drugs" from thirty and forty years ago which frankly are not that good. But it's hard to reform because the fraction of people who ever think about cancer as a problem to be solved is quite small; and it ought to be far larger, given that cancer is the second or even first leading cause of death across much of the world. I wish Elliot Hershberg every success.
I'm pretty much a pessimest when it comes to fighting cancer. I think it's just one of the bugs in our genetic code that evolution didn't shake out. I say that not as a biologist or anyone who has done any work in the field. But I've seen people close to me die of cancer and it seems like the treatement is almost worse than the disease. I agree that the standard first attacks are very crude and have broad systemic side effects and the attitude seems to be "you'll die without this so that doesn't matter."
I read some stuff about mRNA treatment a while ago that seemed like it might be promising.
> seems like the treatement is almost worse than the disease.
I think that's what the poster above you was saying. "Oldschool" chemo is basically poison, and the hope is that it kills off the cancer before the patient. But there are newer drugs that are extremely effective with way way way less side effects out there, depending on which type of cancer one has. Things like immunotherapy are really effective if you happen to match their targeted types of cancer, and some have basically 0 side effects, leading to a QoL improvement if they happen to work. People have gotten nobel prizes for some of these discoveries, it's really insane how far we've come in the last 30-40 years.
> I think it's just one of the bugs in our genetic code that evolution didn't shake out. I say that not as a biologist or anyone who has done any work in the field.
I'm just curious, do you know what the opinions about this stuff are from people that work in these fields, or that have dedicated their lives to it?
I work in this field. It’s more or less correct but kind of lacking in detail. Cancer is a property of all multicellular life. I think it’s best understood as the behavior of a dynamical system that loses the feedback control that keeps cell growth under control.
It’s a bit jargon heavy but it’s a nice case study in how tumor growth is controlled through all the same mechanisms that normal tissue growth uses. Even cells with an outright cancerous gene mutation are basically still just doing normal growth and development.
" Cancer is a property of all multicellular life."
In practice, though, some species are way less prone to cancer than others. Orders of magnitude of a difference, even in mammals. Bats, notoriously. Or naked mole rats. On the other hand, mice get cancer fairly reliably.
Which means that there are biologically realistic way how to keep the danger at bay, and they seem to involve the immune system.
I am guessing: There is an evolutionary "shadow". Genes for getting old and healthy are not selected for, because you get old after having children. Evolution optimizes for the survival of your children.
Might be that cancer hits after creating offspring.
In a social species such as ours, with such a prolonged childhood, having healthy parents and grandparents is likely to affect the survival of children so there will be some selection pressure on a long life there.
Cancer is too broad of a term.
Some cancers like Hodgkin's lymphoma or testicular cancer respond extremely well to treatment.
Some cancers are caused by cell damage from viruses such as HPV and can be prevented by vaccines.
I get the pessimism because "curing cancer" can essentially be interpreted as "curing aging" but progress is being made.
i'm pretty much a pessimest [sic] when it comes to fighting smallpox. i think it just exploits one of the bugs in our genetic code that evolution didn't shake out.
You can’t say that. Trump is very inconsistent and a consummated liar, so plenty of people didn’t believe on his promises to deliver fascism. And plenty of people did believe on his promise to end wars. /s
Whether your little black heart wishes concentration camps or you’re just hoping your paycheck goes a bit further, voting for a con man is a terrible idea.
Yes. The problem is the level of the conversations: "AI is good for this. It's not good. It gaslights." It doesn't come out of that. If we were talking about the actual layers in the models and how they interconnect, at least there would be quite a bit of variety in the conversations. Also conversations about how is all about to end tend to be fun if the interlocutors are creative enough.
But the sooner we get to the part of history with the chromy-killer-robots and people-sabotaging-datacenters-and-foundries, the sooner we will get some meaningful excitement.
The solution for "Windows" is to use the current US derangement to convince the EU, China and even the Russians that, as a matter of national security, it's in their best interest to come to the negotiation table and create an international organization that "adopts" WineHQ[^1] and any other projects that focus on a Win32 compatibility layer. It can be financed with increased sales taxes for any PC software distributor that doesn't natively support Linux.
Yes and no. Putting Sam Altman, OpenAI and even LLMs in general aside for a moment, one doesn't need to look too hard to uncover the failure modes of human minds when it comes to research. Here are some from my personal experience:
- mathematical complexity. This hits hard anybody wanting to understand quantum mechanics, even the parts of it which are many decades old. But one is not going to be making a great many things without that understanding. Right, one may come back and say "but who needs new toys? Let's all go to Church instead!" which is indeed the only valid counterargument one can make, read below if you don't believe me.
- biological complexity. One needs a needle smaller than the size used to pin Drosophila flies to poke a biologist and get them to raise their hands in the air and say "it's all too complex. Nobody can really understand all of that. We know nothing of those systems. We are but humble observers trying to grasp at straws". If said biologists really wanted to understand everything, the majority of them would follow the tiny minority who are trying to simulate biological systems at a quantum level, but see my point above. And that simulation effort is perhaps even more complicated than the biological systems under study, even if it has the potential to scale orders of magnitude better.
Without any of it, cancer will get one in eight people before the age of 65, and ageing and other diseases will eat the rest of us. Which is okay of one believes in God, ofc. Sam Altman and his chummies probably don't believe in the Almighty. In so far as I'm concerned, if they want to fuck cancer, they have my blessing to boil the oceans.
Genes of the wild cabbage: "yah man, we will turn this leafy body into whatever you like. That you are going to eat it? We don't mind a bit, as long as you make more copies of us; that's all that matters."
It's not looking good, I don't think supply is catching with demand yet.
Though the other day I learned there are many technologies for "RAM", and most of them are garbage for LLMs but still useful for other things, like microcontrollers. So I'm thinking my next "build" is going to be a guitar.
> Though the other day I learned there are many technologies for "RAM"
I'm an advocate of sticking a $5 16Gb Optane stick from eBay on a $10 M.2 to PCIe 1x adapter from eBay. Set it up as swap in Windows or Linux. Or pay $200 for a 16GB stick of DDR5.
> I don't think supply is catching with demand yet.
It will take a couple years. Either AI has to have a big crash or YMTC has to grow their production 3-8x before prices come back down to levels from 12 months ago.
Super interesting charts there. What's really interesting to me is that the GPU prices (which also includes RAM) didn't see such a massive increase in price as the RAM itself. Anyone know why that is?
I held my nose and bought an RTX 5070 Ti for $100 over MSRP in January. The very next week the same model was up $200. It turns out that NVIDIA had been subsidizing retail graphics cards with its Open Pricing Program. Not the whole story, but it may help explain the relative flatness of the graph until the end of January.
The other part of it is that the MSRP already baked in a substantial increase from the previous generation. While RAM was near rock-bottom pricing when this hit, current-gen GPUs definitely were not.
A $1500 5800 only has 16GB which would be $250 if you compare it against the DDR6 graph on that page. Given that there's only 2 top tier GPU manufacturers at most, they were probably already not very BOM cost sensitive.
> The company would compensate them well: full wages for the whole time, even when sleeping, and a paid week off after.
Only a week??? I mean, we are all going to be replaced with AI any day now, but were it not the case, I'm fully expecting to see an American company to offer, as a benefit, "we will collect and bring your remains to the workplace if you by accident die outside."
reply