storus

28 ms·

511.
by storus 1y ago
LLaMA 4 is barely better than LLaMA 3.3 so a year of development didn't bring any worthy gains for Meta, and execs are likely panicking in order not to slip further given what even a resource-constrained DeepSeek did to them.
512.
by storus 1y ago
Let's pretend BitKeeper didn't exist. Also, that's one way to misinterpret what was said. I acknowledge there are some well-designed pieces of software that were hacked quickly. LangChain is not one of them, it's a mess,
513.
by storus 1y ago
It's on langchain's official page, a framework that looks like it was hacked over the weekend by a fresh grad that brought a lot of pain to the agentic development, and this just feels like piling up more pain on it.
514.
by storus 1y ago
"I hacked on an open source package (deepagents) over the weekend." Thanks but no thanks.
515.
by storus 1y ago
I would look at PQQ, CoQ10, B-complex, GlyNAC or just glycine, AXA1125, R ALA, DCA, creatine; those are known to improve mitochondrial fitness under various mechanisms. Add 99%-100% dark chocolate and exercise, both of which act similarly t
516.
by storus 1y ago
Humming increases NO as well. Maybe you should hum after taking Listerine for a while? I thought one of Listerine's main (unwritten) selling points was that it helps with sex (smell, prophylaxis for STIs).
517.
by storus 1y ago
You are probably unaware of unintended consequences enabled by this app - many women use it to find bad boys they feel attracted to due to some brokenness in female psyche. So you'll get public outrage on one hand and private DMs on th
518.
by storus 1y ago
Are you libertine's alter ego? Reusing the same arsenal of logical fallacies? Don't you feel like you should be original and forge your own path through these dark arts?
519.
by storus 1y ago
Could this come from the hypothesis that covid could be oncogenic? https://pmc.ncbi.nlm.nih.gov/articles/PMC10202899/
520.
by storus 1y ago
I listed all logical fallacies and "schwarze rhetorik" kinds I could find in your previous statements and I am mighty impressed you fitted so many in such a short space. I was wondering if you are a student of these dark arts or i
521.
by storus 1y ago
Strawman fallacy, poisoning the well, circumstantial ad hominem, implied tu quoque, accusation & moralistic framing, dismissive language & loaded terms, shifting blame & victim reversal, false attribution of motive, appeal to ir
522.
by storus 1y ago
AFAIK X is owned by an American billionaire, not a Russian/Iranian one, so please don't pollute the argument with irrelevant strawmen. What we are seeing real-time is a failure/hack of representative democracy that serves onl
523.
by storus 1y ago
That's becoming the new "think of the children".
524.
by storus 1y ago
Why should the billionaires that set the current EU social networking policy be more important than the lone billionaire that runs X? Or do you think these policies are organic and come from the EU population itself?
525.
by storus 1y ago
OK, so EU is allowed to manipulate social media under various threats, and now when some social media says enough of this, they are criminalized? And people are applauding it?
526.
by storus 1y ago
Just today I ordered a 32dB Liebherr; the previous one had 35dB and could be heard all around the studio (I measured the noise using a dedicated sound meter).
527.
by storus 1y ago
Isn't Mistral based on LLaMA, a US model?
528.
by storus 1y ago
Maybe there is a reason why is your brain hesitating (e.g. low energy) and pushing it with LLMs would blow it up at some point in the future if the cause was not fixed in the meantime? Covid showed me (in a very accentuated way) that many o
529.
by storus 1y ago
So we have another ugly VSCode clone with the same legal/support issues as all the other clones, but without any vibe-coding worthy assists like voice dictation/control or any other major differentiator. Just another "me too&
530.
by storus 1y ago
Doubling of the prices overnight coming like in Croatia and many countries before it?
531.
by storus 1y ago
I think it feels way more realistic than Jurassic Park. Unbelievable what they were able to do back then. https://www.imdb.com/title/tt0047930 https://www.youtube.com/watch?v=_xBl_FiXG8Q
532.
by storus 1y ago
Can Mercury use tools? I haven't seen it described anywhere. How about streaming with tools?
533.
by storus 1y ago
If your pain stems from covid, try high doses of thiamine or alternatively TTFD. Quite a few people reported feeling great for the first time in ages after a single high dose and there seem to be some studies showing pain disappearance in f
534.
by storus 1y ago
Blood tests are often misleading, not showing tissue deficiencies. For those there might be 1-2 labs in the world that can do them depending on the type of deficiency.
535.
by storus 1y ago
My guess is the main issue is latency and accuracy; a single agent without all the routing/evaluation sub-agents around it that introduce cumulative errors, lead to infinite loops and slow it down would likely be much faster, accurate
536.
by storus 1y ago
My point is that newer models will have those baked in, so instead of supporting ~30 tools before falling apart they will reliably support 10,000 tools defined in their context. That alone would dramatically change the need for more than on
537.
by storus 1y ago
You can look at AnyTool, 2024 (16,000 tools) and start looking at newer research from there. https://arxiv.org/abs/2402.04253 For long contexts start with activation beacons and RoPE scaling.
538.
by storus 1y ago
Those issues are considered artifacts of the current crop of LLMs in academic circles; there is already research allowing LLMs to use millions of different tools at the same time, and stable long contexts, likely reducing the amount of agen
539.
by storus 1y ago
Isn't it due to synaptic pruning going wrong, a sort of off-by-one error of a gene tagging synapses for deletion?
540.
by storus 1y ago
You need at least 5x Pro 6000 (for smaller contexts), let's say Max-Q edition running at 300W, so overall you get a minimum of 1500W. You get around 6 tokens/second which is not great but not terrible. If you use very long prompts

More ›