jes5199

a meme I heard 15 years ago is:

What’s the definition of “Legacy Code”? It’s code that has business value.

(otherwise you’d just delete it and move on)

Tony himself doesn’t post frequently, but his new book “Stellar” is worth checking out. Some of the other RethinkX people, Adam Dorr @adam_dorr and Bradd Libby @bradd_libby are better at posting high-frequency updates using Seba’s models. former-RethinkX researcher Nafeez Ahmed has a blog that is also worth checking out: ageoftransformation.org

I've been following Seba since around 2020, and it's wild how: 1) he continues to be correct 2) people are still very slow to believe him

his predictions have gotten a LOT more dramatic lately, I can't wait to see if he's still nailing it

here you go: https://www.tonyseba.com/wp-content/uploads/2020/11/Rethinki...

I vote blue but I don’t see the dems fast-tracking fission! Or really anything else particularly effective about climate for the last 30 years.

anyway my model is: we just have to survive a few years before Wright’s Law pushes solar+batteries so cheap that fossil is priced out of the market. Thus aerosol injection to bridge the gap until drawdown

yeah I tried to build a project on Automerge but I ended up switching to Yjs, it seems more mature.

no, just some acquaintances. but the only people even doing the experiment, as far as I can tell

with Intel also shitting the bed, it seems like AMD is poised to pick up “traditional computing” while everybody else runs off to chase the new gold rush. Presumably there’s still some money in desktops and gaming rigs?

great, let’s scale up https://makesunsets.com/

good survey of what people are already implementing, but I’ve convinced we barely understand the possibility space here. There may be much more elaborate structures that we will put context into that haven’t been discovered yet

i’d really prefer a timeline more specific than “in the long term”

okay then why is it taboo to suggest geoengineering interventions like injecting sulfer into the upper atmosphere? The climate advocates don’t have any decisive and ambitious actions that they actually are willing to try.

analogy: a CPU is more expensive, more complicated, more energy demanding than custom made circuitry, in most cases.

oh I have use of that, never heard of it

okay I’m practicing my new spiel:

this focus on coding is the wrong level of abstraction

coding is no longer the problem. the problem is getting the right context to the coding agent. this is much, much harder

“vibe coding” is the new “horseless carriage”

the job of the human engineer is “context wrangling”

postrats were never a coherent group but a lot of people who are at https://vibe.camp this weekend probably identify with the label. some of us are still on twitter/X

I haven’t tried pro yet but just yesterday I asked O3 to review a file and I saw a message in the chain-of-thought like “it’s going to be hard to give a comprehensive answer within the time limit” so now I’m tempted

this is unhinged right? like what on earth does starlink have to do with voting machines?? how is this conspiracy supposed to work, on a technical level

I think the Apple paper is practically a hack job - the problem was set up in such a way that the reasoning models must do all of their reasoning before outputting any of their results. Imagine a human trying to solve something this way: you’d have to either memorize the entire answer before speaking or come up with a simple pattern you could do while reciting that takes significantly less brainpower - and past a certain size/complexity, it would be impossible.

And this isn’t how LLMs are used in practice! Actual agents do a thinking/reasoning cycle after each tool-use call. And I guarantee even these 6-month-old models could do significantly better if a researcher followed best practices.

so this did all this transparency stuff without it being the run up to the launch of a new augmented-reality device? I don't understand what they are thinking.

I lost the joy of coding years ago. Using an LLM is novel and interesting, in way that programming hasn't been for over a decade

if it’s gentle, it’s not a singularity

I’m not sure we’ll be metering electricity if Wright’s Law continues

> Like most employees I blamed myself for not working hard enough to get good compensation

ohh. I feel like I understanding something about my peers now that I had not caught onto before

avoiding monetary control policies?

I tried Cline, but I like Cursor better

there's a myth I know where the true Forth enthusiasts end up making their programs so pure that they start needing their own special hardware (in one version, I heard that machine words were all 13 bits), their own networking stack, their own communication apps. Until eventually they transcend into a second internet, populated only by forth hackers, where everything is optimally elegant.

and that's why we don't hear from them anymore

also, the agents are actually pretty good at cleaning up spaghetti if you do it one module at a time, use unit tests. And some of the models are smart enough to suggest good organization schemes!

there will be _less_ fossil fuel use, globally, in 2030 than there is today, and AI will not be a significant reason that changes in either direction.

the “environmental costs” argument is based upon like fifteen false assumptions. by 2030, AI will be running almost entirely on solar+wind+battery, because that is by far the cheapest option