worst scenario for you is most likely - MSTR blow up, BTC bottom and recovers, while MSTR fade away
if you want leveraged bet on BTC - by BTC, use it as collat on lending, buy more, keep adding on lower prices. It will hardly be -50% from here, most likely 30% is maximum downside, so you have pretty much room to do x2.5-x3 looped
that's definitely is true, hence USA is in top rankings. but it's tradeoff - you can't have pros of having guns without cons
also, if guns are acquired illegally, it might be harder to identify shooters. still order of magnitude higher than in other first world countries, but nowhere as bad as it's shown in raw numbers
this is inacurate, because only like small percent of that is what you meant by gun deaths. most of it is suicide by gun. obviously, in country where people have guns, suicide gonna happen by gun.
anyway, shooting in US is horribly lot in absolute numbers, multiple than in other rich countries.
but for fairness - most of homicide is happening via illegal weapons, by people who obtained that guns illegally, so that's kinda not direct correlation to legal status of guns
you are absolutely missing the point
whole reason of powerful local models is not fast interference or cost effectiveness. With powerful local models one can train and fine tune his own fast models, solving his own tasks on much better speed. Training models, adapting and making them personalized is unique reason, why one needs powerful local models. Fable proved it - they put guardrails on AI development, not only for anti-hackers/malicious actors, but to kill potential competition or do not let local, personal models thrive on their meter. Sooner or later, that guardrails will be everywhere and you'll have no choice to train your own model and be free and forever gonna pay for intelligence to third parties, even if it's cheap. Debate is similar to buying/renting. Obviously renting was cheaper and you could've have nicer house, but building/buying your own house proves itself to be key, as during time it will be impossible. I have local model on my phone and i will have frontier personal models on my machines, but for that reason i need to have tool powerful enough to work on that. Never gonna use commercial GPT/Claude for my personal life and never gonna use my personal local models for commercial things, instead of Claude/GPT.
that's probably wildest pilot i've seen ever from Generative AI (imagegen) to medical tech, doing one of most important breakthrough in modern medicine
Midjourney, the company behind the popular AI image generator, has unveiled what it says is the world’s first full-body ultrasound CT scanner.
>The technology aims to make full-body imaging safer and more affordable.
>During the scan, people are lowered into water while ultrasound waves create detailed 3D images of more than 25 organs and body structures in about one minute.
>The scan uses no radiation.
Midjourney is working toward FDA approval for future diagnostic uses and plans to bring the scanner to market by the end of 2027.
you are completely right and completely wrong at same time
Fusion was our edge for months now, using things like Codex Orchestrator or Gemini Orchestrator or Trifecta (Kimi+Opus+GPT) and it's good that there is broader product now
on another hand, all frontier labs are MoE, which means they already have this internal dialogue in chain of thoughts. Obviously, due to nature of models and them being different NOW, combining them is huge deal. It would not be huge deal, when each of them is super good enough at everything, like Fable and beyond. So in reality, this is more about your toolkit and adjusting things. IDK why you think this is novel idea. It's just never packaged, because that was internal toolkit. There is no FULLSTACK programming language, combining Rust and React, not because it's hard or impossible, just because market don't need it. You want X - you pick A, you want Y - you pick B. rinse and repeat as long as X is not good at A and B.
also, this seems like leveraged bet on non-hallucinating - removing HITL/HOTL from process of inter-LLM dialogue is horrible for production and good for non-practical research toys.
@SamuraiTakedown are you retarded? bear market bottom is around the corner (give it 3-4 months) and after that we will have another almost year of bear, but past bottom. it's like saying you climb up the mountain and then after that top you just go away. you need to climb down..
@frankdegods something like this is built by @UnanimousAI , called Thinkscpace @frankdegods ,check out. you know them, back in a day (2022) they did Cryptoswarm, which failed , but technology moved on so quickly afterwards, worth second look