search_facility

Nothing will happen, we already have kind of translation activities happen between "incompatible views" in real life, it works.

Besides, in weights there are no real incompatibility can happen, all is unified with same math chain through architecture. Different set of local minimums/maximums can have different strong/weak consequences but nothing incompatible.

Different consistent representations also mean that model probably can be simplier for the same task, sign of excessive capacity, imho

Very educative, thanks for sharing the story!

They are used a lot, and modern edit models kind of support base cases - depth, canny, openpose - as a reference image. I mean they understand "use image 2 as depth map", result indeed have same depth, etc. But interpretation sometimes arbitrary to model.

For the tasks where several control nets should be adhered (depth and canny and segmentation) dreaded SDXL is still unmatched. SDXL also the only one handling influence weight for each control net normally, this ability completely vanished on modern models

Control nets in early post-SDXL models like ZIT or Flux kind of work too, but still much worse in quality when combined. I suspect pure diffusion architecture of SDXL is really favorable to mixing control signals with precision, DITs probably can be trained to similar level for mixing signals - but it is much harder and never seen in practice. At least every post-SDXL model I tried is trained with single one in mind, judging from my attempts. This rant is from personal expirience.

Great explanation, thanks! Would be cool to see this for people BSky instead of ex-Twitter

To understand hate against Adobe one have to experience attempt to stop Adobe subscription from renewing. This is truly life-changing experience.

Drone units of occupation forces of Russia are primary taegets for AFU and due limited control distance for FPVs, Russian drone units are serving at the zero line, actually. So they are sent to actual front lines constantly (despite what Russia officially claims)

May be Ethereum should fund 10000 agents to pursue new crypto math with provable strongness/hardness

Also Krita

Yes, why not. "If then ..." Is for "harness", but the answer to prepared question is for jev-like cheap and fast service

Worth to mention that with Claude and GPT this can be result of tournament sampling, which is part of text watermarking. Same answer for all Claude models kind of confirm it, imho.

So not something internal to model thinking.

> AGI is here

If AGI is "Attractions to Get Investments" then yes, it's happening

Well said!

Great, this is unique trait and helpful one - Firefox is here to stay

This "hidden knowledge" is fun to read but pain to remember and use, especially when working with multi-platform environments //

Switched from bash to plain python scripts for shell stuff everywhere several years ago, and never looked back into bash zoo anymore. Stable syntax across Win/Mac/Linux, no bash/zsh/msys2 obscure differences, normal errors and Clause writes scaffolds quick and flawless anyway

... Today its "Claude just needs to write it" and the "hard part" is the prompt :)

Current situation exist presisely due researches of 20th century though. This is largely connected things

There are still problems though - in Chrome translation addon simply no way to login properly so its broken //

Also represents good test coverage

It's the opposite, engineers do know. Claiming otherwise is way too generous and over confident.

I mean between this two "knowings" the Claude inner workings are much more clear for engineers, including many side effects, alternatives, custom shortcuts in processing etc. It's a magic only for people looking at it as black box

Well, at the very least we have a base to distinct different kinds then.

Probably the main problem of people implying LLM consciousness is that they imply LLM have human-kind of consciousness. Judging only on "how they speak", generally, they insist on using the same word that labels human consciousness (exclusively), etc.

But there are so many instrinic differences that such claim is not feasible despite similar "talking abilities".

It is useful, so yes

MOE basically work that way already, QWEN/etc with low active params (A-number in name) allows to inference big models locally (only active params have to fit into memory)

With both cases the agent doing something is human. And human beings are indeed conscious. Outside human needs LLM are useless.

Math, as a tool, is just a proxy for people using LLMs, as well as GPUs spending cycles on calculating the math

> weak reasoning once you get off the beaten path

Yep. And LLM engeneers improving this issues see perfect correlation with only one thing - data quality and quantity through training pipeline. LLM internals are secondary on many metrics for improving that

Humanity just reached the point where collective accessible knowledge covers semi-full perturbations of all main concepts that human consiousness ever produced, with additional associative expanding (math handles this). Full perturbations with current communication complexity are written down and recorded one way or another, LLMs just capitalizing on that tipping point, imho

Math used in LLMs is perfectly reducible and Gödel have nothing to do with it - inside commonly used axioms (which sufficient for LLM to exist and outside of Kurt Gödel scope) there are ZERO questions/uncertainties how it works, it's just a fact :)

+100, companies certainly have direct interest in pumping asset evaluation, and emotional attachment is financial valuable thing. Emotional attachment sells better than xxx this days

Imho no, math itself have no conciousness. Quite confidently its a helpful tool that does not act by himself.

I can not simulate my brain, it's a huge stretch to imply this.

But with LLMs - anyone can simulate LLM. LLM can be simulated without any uncertainties in pen and paper and a lot of time. Does it mean that 100 tons of paper plus 100 years of time (numbers are just examples) calculating long formulae makes this pile of paper consiousness? Imho answer is definitive no.

Since the times GPT-2 was reimplemented inside Minecraft - its quite obvious LLMs are just math. Nothing else, by nature. Modern LLMs have the same math as in GPT-2 - just bigger and with extra stuff around - and math is the only area of human knowledge with perfect flawless reductionism, straight to the roots. It was build that way since the beginning, so philosophy have no say in this :) And because of that flawless reductionism, complexity adds nothings to the nature of math things, this is how math working by design - so it can be proven there are no anything like consciousness simply because conciousness was not implented in the first place, only perfect mimicry.

And the real secret is in the data, not math. Math (and LLMs running it through billions of weights) is just a tool.

With current standartization the issue of "page not working on non-Chrome browser" is non-existent. Thanks god nowadays everything (pages) work everywhere in very similar manner, I am using chrome, firefox, safary and opera and have zero problems last 5+ years. Old days are gone.