HackerLangs
TopNewTrendsCommentsPastAskShowJobs

mhog_hn

no profile record

comments

mhog_hn
·16 dni temu·discuss
Any tips for a 2 person European company. One founder based in Amsterdam, the other in London. We are close to setting things up, some input from other founders would be great
mhog_hn
·28 dni temu·discuss
At d5s.tech we are recreating the layers built on top of models, working on dogfooding our own product to run a large chunk of the company.

I feel extremely strongly that a future in which most companies depend on one or two large AI-megacorps is going to lead to excessive rent seeking sooner or later.

I remain positive that the long term steady state will consist of proprietary models, -but- with open source AI models statistically close.

If compute keeps growing the relative cost of training current frontier models will decrease. An open source Fable/Mythos model simply seems inevitable.
mhog_hn
·5 miesięcy temu·discuss
But what if the AI agent has a 5% chance of adding a bug to that feature? Surely before any feature was completely bug free
mhog_hn
·5 miesięcy temu·discuss
It is the same with the global financial system
mhog_hn
·5 miesięcy temu·discuss
For the CLI > MCP folks: https://cliwatch.com/
mhog_hn
·5 miesięcy temu·discuss
https://cliwatch.com/ for any CLI maintainers that aim to keep track of how agent ready their work is :) get in touch! very responsive to feedback
mhog_hn
·5 miesięcy temu·discuss
Asimov’s Foundation series
mhog_hn
·5 miesięcy temu·discuss
Amazing. Would be cool to see agents end up trading at varrock bank like during the old days. Sort of a facebook/moltbook equivalent - wonder how genuine it would feel
mhog_hn
·6 miesięcy temu·discuss
Still in awe that AGI is happening
mhog_hn
·6 miesięcy temu·discuss
“LLMs bailed us out of the impending ultra-specialization” - well said!
mhog_hn
·7 miesięcy temu·discuss
Imagine throwing orders of magnitude more of compute at things - we may have things like a monte carlo tree search for LLM outputs using an LLMJudge that prunes the tree.
mhog_hn
·7 miesięcy temu·discuss
On the one hand it is a wholesome article. On the other hand - so much wasted potential of people squeezing out the last bits when competing. Nash equilibria can suck
mhog_hn
·9 miesięcy temu·discuss
Title could also be “How to train biological neural networks” - Andrej Karpathy
mhog_hn
·10 miesięcy temu·discuss
One general purpose web framework away