guluarte

8 ms·

151.
by guluarte 1y ago
protip: never trust the client
152.
by guluarte 1y ago
this seems like a fake circular economy, ms invests in openai which uses the money in azure, amazon invests in anthropic which pays aws for hardware and infra, nvidia invests in openai which uses the money to buy nvidia hardware, etc
153.
by guluarte 1y ago
no, AI companies need to continue to say things like that and do "safety reports" (the only real danger of an llm is leaking sensitive data to a bad actor) to maintain hype and investment
154.
by guluarte 1y ago
it's a revolutionary app the let's you create calendar events to have dinner with your friends
155.
by guluarte 1y ago
Look at me, you are the model now
156.
by guluarte 1y ago
Robot: $1,500 Refills: $5,000
157.
by guluarte 1y ago
companies are just riding the hype of the last tool, last season was cursor, then clis and now browsers
158.
by guluarte 1y ago
interesting, what if we put datacenters in the ocean floor with nuclear power? like the Army Janus program
159.
by guluarte 1y ago
just use bruno+git https://www.usebruno.com/
160.
by guluarte 1y ago
i just use https://github.com/ravitemer/mcphub.nvim and enable/disable the mcps i want
161.
by guluarte 1y ago
the hardest part of scrapping is bypassing Cloudflare/captchas/fingerprinting etc
162.
by guluarte 1y ago
LLMs will never lead to AGI, never, does a PhD know all the internet? no but he can create new knowledge, LLMs are trained with all the data possible to cover most cases and they are excellent for autocomplete
163.
by guluarte 1y ago
that's probably a good idea, either AI bubble explodes or competitors catch up
164.
by guluarte 1y ago
my bet is we will just slowly automate things more and more until one day someone will point out when we reached "AGI"
165.
Andrej Karpathy – "RL is terrible; everything else is worse" [video](youtube.com)
1 points by guluarte 1y ago|0 comments
166.
by guluarte 1y ago
it also may point out that the solution for context rot may not be coming in the foreseeable future
167.
by guluarte 1y ago
great! another set of files the models will completely ignore like CLAUDE.md
168.
by guluarte 1y ago
any web3 that sends you a test project is a scam and are super common on sites like upwork and linkedin
169.
by guluarte 1y ago
AI is a massive bubble, nvidia invests in openai which buys nvidia chips, nvidia is just doing round-trip transactions
170.
by guluarte 1y ago
I think the reason could also be a lot of countries and companies started to diversify more and depend less on the USA
171.
by guluarte 1y ago
"Hey you #$#@ remove the ear from my anus"
172.
by guluarte 1y ago
another option is to use Prometheus+grafana https://docs.nvidia.com/datacenter/cloud-native/gpu-telemetr...
173.
by guluarte 1y ago
are not mac servers more expensive than traditional ones? the only reason i've used them is to compile xcode
174.
by guluarte 1y ago
social anxiety i think it is more the fear of breaking your own perception of yourself
175.
by guluarte 1y ago
I was using it daily, but after the price jump, using codex and claude was much cheaper than using deepseek.
176.
by guluarte 1y ago
by your logic then you have to disqualify openai and anthropic first party endpoints for testing gpt and claude...
177.
by guluarte 1y ago
not only that, google now has like 10 AI apps, like literally I lost count.
178.
by guluarte 1y ago
markov chains with limited self correction
179.
by guluarte 1y ago
lol a "mistake"
180.
by guluarte 1y ago
Most leetcode problems fall into the same ~15 patterns, and hard problems most of the time require you to use a combination of two patterns to solve them.

More ›