In practice, full access to your machine is okay as long as there are safeguards and the expected outcomes are clear with a well defined path to said outcomes that aren’t overly ambitious. Otherwise, for ambitious goals or YOLO one shot attempts, eliminating opportunity for capability misuse is critical (e.g., sandbox).
Objectively terrible policy for ethics, public safety, and, selfishly, the American economy. Immigrants contribute to economic growth and are less likely to commit crimes are well established facts. It’s the 21st century, we have the internet and education is accessible, but instead of recognizing and championing the vital role of immigrants in America’s rise to power, here the nation moves to hurt itself for some misguided anti immigrant ideology.
A lot of comments here are about the issues with the analysis in OP’s post but much of them are “a distinction without a difference” with respect to the broader conclusion. When we look at purely cost and performance (setting aside privacy) then it’s better for individual devs to pay for hosted then for self hosting. Employers are paying for tokens on the job and most devs are finding the $PREFERRED_PROVIDER’s $20/$100/$200/month subscription sufficient outside of work. Most devs don’t fall in the conditions under which running local models make sense purely on the basis of cost vs performance.
More critically, in practice, setting up local models seems more like a hobby, an educational exercise, or an act of privacy control than it is for cost cutting or productivity.
There’s so many ways, many redundant, to set up agents for software development that beyond personal/team/org needs+tastes, I need to look into setting up some benchmarks to evaluate what set up is optimal or whether the differences are even worth it.
> the NSF funds approximately 25% of all federally supported basic research conducted by the United States' colleges and universities.[5][6] In some fields, such as mathematics, computer science, economics, and the social sciences, the NSF is the major source of federal backing
Thanks! I want to do something similar to your LLM suggestion, the endgame is tooling for forums and individuals to improve the quality of discourse. More broadly, I think LLMs and recent advancements now make it possible to assist with self improvement (e.g., see former startup Humu’s nudges but for everyone instead of just B2B)
> applying this compression algorithm at scale may significantly relax the memory bottleneck issue.
I don’t think they’re going to downsize though, I think the big players are just going to use the freed up memory for more workflows or larger models because the big players want to scale up. It’s a cat and mouse race for the best models.
> You do not have the right to eliminate yourself, you do not belong to you. You belong to the universe. The significance of you will forever remain obscure to you, but you may assume that you are fulfilling your significance if you apply yourself to converting all your experience to highest advantage of others. You and all men are here for the sake of other men.
Let’s say it costs $10K/month/person so $120K/yr/person. Probably a big overestimate but gotta include healthcare and help people with long term stability.
That’s 120,000 x 1,000,000 = 120,000,000,000 or $120 billion USD.
Idk what the Nth order effects would be but yea I think what you’re saying tracks in the numbers
There’s a lot to address here but in short: VFIAX is an index fund, it tracks the S&P500 index, it’s not actively managed, SpaceX will likely be in the S&P500, so my comment around VT applies to VFIAX (as far as the question of exposure is concerned) but to a greater extent than VT (see VT’s composition vs VFIAX’s composition).
Obligatory not financial advice, I’m not an expert, don’t make any financial decisions based on hacker news comments, etc
My understanding: It depends on what index the fund is tracking. QQQ tracks the Nasdaq-100 so QQQ is vulnerable. VT tracks the FTSE Global All Cap Index so VT is not directly affected by Nasdaq’s choices but is still exposed to some extent because spacex is likely going to be in the aforementioned FTSE index, Nasdaq’s actions impact spacex’s market cap, and thus Nasdaq’s actions impact spacex’s position in the aforementioned FTSE index which in turn affects VT’s composition (to a smaller extent than QQQ’s).
EDIT: to be clear the above are just examples with two funds (QQQ and VT)
I work at aws and generally use Claude Opus 4.6 1M with Kiro (aws’s public competitor to Claude Code). My experience is positive. Kiro writes most of my code. My complaints:
1. Degraded quality over longer context window usage. I have to think about managing context and agents instead of focusing solely on the task.
2. It’s slow (when it’s “thinking”). Especially when it’s tasked with something simple (e.g., I could ask Claude Opus to commit code and submit for review but it’s just faster if I run the commands myself and I don’t want to have to think about conditionally switching to Haiku / faster models mid task execution).
3. It often requires a lot of upfront planning and feedback loop set up to the extent that sometimes I wonder if it would’ve been faster if I did it myself.
A smarter model would be great but there are bigger productivity gains to be had with a good set up, a faster model, and abstracting away the need to think about agents or context usage. I’m still figuring out a good set up. Something with the speed of Haiku with the reasoning of Opus without the overhead of having to think about the management of agents or context would be sweet.