spongebobstoes

6 ms·

61.
by spongebobstoes 4mo ago
copyright law stops illegal copying you are worrying about a hypothetical without precedent. all past evidence is that new creative tools are good
62.
by spongebobstoes 4mo ago
this is the same argument against the printing press, photography, home printers, youtube, etc it is a bad argument. more accessible tools consistently lead to more creation in a positive way
63.
by spongebobstoes 4mo ago
because it's a dead end of the technology. reasoning will be everywhere soon enough, even for free accounts to know what most people experience, this research should be based on the in-app experience. it isn't, so it misses that m
64.
by spongebobstoes 4mo ago
discovery has always been the problem with art. many people are instrinsically motivated to create, that's why we have so much art AI is orthogonal to human connection. people like people
65.
by spongebobstoes 4mo ago
this has reasoning disabled everywhere, making it a pretty bad benchmark. the argument given is that's the "default consumer experience" that might be generally true, but I think chatgpt has reasoning enabled for free account
66.
by spongebobstoes 4mo ago
they just want data centers now. most companies would rather use solar, but they can't on short timelines due to land use regulations (and import tariffs)
67.
by spongebobstoes 4mo ago
personal computing is inseparable from technological progress, and has been wildly positive for humanity
68.
by spongebobstoes 4mo ago
transparency is knowledge, and is a prerequisite for accountability
69.
by spongebobstoes 4mo ago
my take is simply three points 1/ telemetry is operationally useful in systems engineering 2/ the human body is a system 3/ this is a step towards telemetry for the human body > currently we have almost no data this is
70.
by spongebobstoes 4mo ago
this doesn't need to be a diagnostic model, just a data source for existing doctors
71.
by spongebobstoes 4mo ago
"the people can't be trusted with knowledge" is a bad argument we don't need to do much differently to take advantage of this data anyway. doctors already ask patients what changed recently collect data passively. when a
72.
by spongebobstoes 4mo ago
those scans are one-off events, these scans are regular the signal is improved by focusing on differences over time, instead of looking for insight from a single snapshot in a production system, I look at the change log around incident star
73.
by spongebobstoes 4mo ago
do they need FDA approval to have a good impact? it doesn't need to be a licensed medical device for me to interpret the data, and then verbally tell my doctor what's wrong this is similar to how people get a lot of medical value
74.
by spongebobstoes 4mo ago
they aren't suggesting it as an actual strategy, merely an indicator that the fundamentals seem realistic
75.
by spongebobstoes 4mo ago
it's completely false
76.
by spongebobstoes 4mo ago
they have 6B left over after paying for inference, that's a lot of money R&D is a leading expense, a good portion of that is probably R&D for 2026 models
77.
by spongebobstoes 4mo ago
this just isn't true, openai is largely (probably 80%+) consumer revenue in 2025
78.
by spongebobstoes 4mo ago
because they have more research and better models coming building a Rube Goldberg machine on Chinese models might work okay, but it will be brittle, and is unlikely to work as well as the latest and greatest model from OpenAI the demand for
79.
by spongebobstoes 4mo ago
all the best coders and engineers I've worked with have been nerds and very passionate about computers passion is necessary but not sufficient to be the best. you can be top 20% without passion, but you can't be top 1%
80.
by spongebobstoes 4mo ago
I think the overly sensitive filters reveals an alignment gap if they had more success on alignment and safety research then I don't think the cludgy filters would be necessary
81.
by spongebobstoes 4mo ago
this is just not correct. the brightdata link is a data set scraped from Facebook, not sold by Facebook Cambridge Analytica was the same thing, scraped from Facebook by people who voluntarily (perhaps ignorantly) gave their data to quiz app
82.
by spongebobstoes 4mo ago
no, it doesn't sell user data. or, where can I buy some?
83.
by spongebobstoes 4mo ago
arena is not a good benchmark, it is very susceptible to sycophancy
84.
by spongebobstoes 4mo ago
he pushes mysticism of the models he's starkly anti-China with a warlike posture that I find dangerous and unappealing Anthropic has a much more confused mission statement than OpenAI in interviews, Dario appears to care little for the
85.
by spongebobstoes 5mo ago
this essay is asking the right questions
86.
by spongebobstoes 5mo ago
I don't think it is required, democracies aren't designed that way not everyone in charge is a cartoon villain
87.
by spongebobstoes 5mo ago
this whole blog post is basically "people need jobs to be happy, so we should design our society such that they need jobs" not only is the premise wrong, but forcing people to work is not a good or ethical way to address this prob
88.
by spongebobstoes 5mo ago
with revenue >2x cost, they can afford to have a miss now and then
89.
by spongebobstoes 5mo ago
each new model brings in revenue that is multiple times the cost to create said model
90.
by spongebobstoes 5mo ago
I think we might just disagree, and that's okay the effectiveness of online ad targeting enables a lot more businesses to be profitable, this is indisputable it is still a choice to say "I need to run a visual art business that ma

More ›