SailingSperm

Love the idea - wonder if it can be worked directly in with a system prompt like 'write <bkspc> to nullify the closest preceding non-nullified word'. Then just parseing in the output. eg. "hi how are <bkspc> tall are <bkspc> <bkspc> do I talk" -- would resolve to "hi how do I talk"... Or maybe better with <bkspc[index#backFromThisPosition]>

Can't think of a good query to test it on that might need revision half way through...If anyone's got some ideas?

I do find with some coding problems LLMs can start a solution then as it describes its solution it needs to contradict what it has said earlier from it providing/working out itself more context to the problem.

Seems a little sensitive... Prompt: "Taiwan" === ''' This prompt is being reviewed We're taking a closer look to make sure this prompt doesn't conflict with our content policy. '''

Interesting getting that terse response from gtp4. I got a seemingly far more considered and practical response from 3.5...Seems it may be that gtp4 is accounting better for the light hearted / non-seriousness of the question and responding in turn with a fairly non-serious resp.

3.5-turbo:

As a language model, I cannot say for certain how a whale would perceive or understand the concept of being wet. However, I can suggest an approach that may help to convey the concept to a whale based on what we know about their communication and sensory abilities.

Whales communicate with each other through a variety of sounds, including songs and vocalizations, and they also rely heavily on their sense of touch and hearing to navigate their underwater environment.

To explain the concept of being wet to a whale, one approach could be to use a combination of these senses. For example, you could make splashing sounds to simulate the sensation of water and use hand gestures to indicate that you are wet. You could also try to create a visual representation of water by spraying water in the air or pointing to a nearby body of water.

Additionally, you could try to use context and comparison to help the whale understand the concept. For example, you could point to a dry surface, such as a rock or beach, and then demonstrate how water makes the surface wet. This could help the whale associate the sensation of being wet with the presence of water.

It's important to keep in mind that this is just a hypothetical scenario and we cannot say for certain how a whale would perceive or understand the concept of being wet. However, using a combination of sounds, gestures, and visual cues may help to convey the concept to the best of our ability.

They're saying there's a distinction of being able to communicate freely (and possibly lie) while still not showing your final action to defect or cooperate. Which yes, leaves the prisoners dilemma much the same as the 0 forms of comms version.

They may be talking about immunological imprinting[1] - Something possibly resulting in a less than optimal immune response - from the initial vaccine creating highly specific antigens, thus not as comprehensive as an initial natural infection might give to prime the immune system against the virus and its variants. [1] https://en.wikipedia.org//wiki/Original_antigenic_sin

>the advocacy of a small government is exactly _more_ situations like FTX. Arguably not true. The presence of large government and pervasive regulation may behave like the bike helmet paradox (Where cyclists feel they are able to take more risk while wearing a helmet, and motorists drive more carefully around helmet-less riders). Smaller gov in this case may lead to individuals feeling less like the government has vetted xyz investment opportunity and therefore may feel a need to do their own due diligence more thoroughly.

>Can you explain how this standard of rigueur leads to anything less than a complete halting of medical research?

~10 yrs of trials and safety data collecting (Including long term studies) is the historical norm. We've certainly not had a complete halting of medical research from doing that.

My main point though is that with the level of data provided and the fact long term studies have not, and cannot have been conducted in this time frame - the push to mandate (and coercively pressure vaccination uptake through fear of losing your job, or access to regular life things bars, cafes, travel etc.) is unjustified and arguably doesn't even reach the bar for informed consent (with many people getting it under the impression it would prevent contraction and transmission).

Though, firing medical staff based on vaccine status(that has little impact on transmission) while saying "we need all the medical staff we can get to ensure healthcare facilities don't collapse" doesn't square up. Here in Australia it's still the case that our hospital staff require the vax, while we're deeply unstaffed. It begs the question if their exclusion is of any net gain to society despite the ostensible risk they pose to patients, or if their ability to contribute to the healthcare workforce would outweigh that 'risk'.

The 12 billion or so COVID vaccine doses given also only have a history of ~2 years. If the adverse effects of it were say for instance, minor cardiac damage that resulted in no acute symptoms, but instead an overall reduction in life expectancy by 5 years, the costs of such would not be known for some time. Another side of this is the possible cumulative effects of ongoing covid vaccination boosters that have no long term data.

Finally also the danger of immunological imprinting[1] possibly resulting in less than optimal immune response. The fact these vaccines have little impact on transmissibility means basic evolutional biology theory is at play, driving selective pressures for antigens humans are less capable of mounting defenses against thanks to an already primed imuno response of the original virus' antigens. This is evident with the proliferation of variants and sub variants.

[1] https://en.wikipedia.org//wiki/Original_antigenic_sin

Check out the free, open source https://stablehorde.net/ for stable diffusion image generation via a distributed cluster (with REST API).

A Crowdsourced distributed cluster of Stable Diffusion workers. With amazing parallelism opportunities.

fully documented REST API: https://stablehorde.net/api

Client made with Godot.

Also an installation-free client: (run in web or download)

https://dbzer0.itch.io/stable-horde-client

Free to use - but 'joining the horde' with your gpu will get you 'kudos', increasing your priority when generating.

Credits Db0 ( https://dbzer0.com/ )

Reddit thread: [Stable Horde, the crowdsourced SD API, has recently passed 1 gigapixelstep of generated images in 75K requests, over slightly less than 2 weeks.] https://www.reddit.com/r/StableDiffusion/comments/xrxoxo/sta...

A little Titanic factoid: The novella "Wreck of the Titan" written by Morgan Robertson and published as Futility in 1898, and revised as The Wreck of the Titan in 1912. It features a fictional British ocean liner Titan that sinks in the North Atlantic after striking an iceberg. The Titan and its sinking are famous for similarities to the passenger ship RMS Titanic and its sinking 14 years later.

"Biggest ship in the world" title

Written before the RMS Titanic was even conceptualized

The fictional ship sank in April in the North Atlantic 400 miles far away from Newfoundland, and there were not enough lifeboats for all the passengers. (all same as Titanic)

https://en.wikipedia.org/wiki/The_Wreck_of_the_Titan:_Or,_Fu...

And they probably do the same thing as 5 eyes[1] to 'stay honest' - have their allies do those things to their citizens and them do it to their allies, then share the intel.

[1] https://en.wikipedia.org/w/index.php?title=Five_Eyes#Domesti... (a treaty for joint cooperation in signals intelligence.)

Nice, the inbuilt levels editor after they're generated is a nice touch.

I think it's actually the opposite. shrink the img and give it a white border, then re-imagine. Stitching together and playing the opposite way to give the zoom effect. shrink to keep same px dimensions or just enlarge canvas and place img in center.

Some good examples of this being done with Dalle2 last month -- https://youtu.be/TW2w-z0UtQU?t=244

For the fungus gnats -- Try 'Mosquito Bits' a biological way, using bacteria, of killing them while they're larvae in the soil-- Bti(Bacillus thuringiensis israelensis) produces toxins which are effective in killing various species of mosquitoes, fungus gnats, and blackflies, while having almost no effect on other organisms. They're very effective.

Interestingly, it was first found in a stagnant pool of water in the Negev Desert. (perhaps the only place found to naturally occur)

https://en.wikipedia.org/wiki/Bacillus_thuringiensis_israele...

Had something maybe similar too - some times instant connecting, sometimes slow, sometimes 0%. Fixed and happily using ZT every day now. (TY ZT!)

On both computers do: Network connections->Right click ZT network->Properties->Select (TCP/IPv4), probs 5th down. ->Properties->Advanced->Untick automatic metric-> set interface metric value to 1. Potentially reboot, though I've not needed to before.

Just for some visuals if needed https://support.parsec.app/hc/en-us/articles/115002766652-Us...

Apologies! I misread your comment - which I'm not sure how, upon re-reading and finding it pretty clear.

>Adam Smith's famous "It is not from the benevolence of the butcher, the brewer, or the baker that we expect our dinner, but from their regard to their own self-interest" thesis is wrong because shopkeepers of the time mostly sold goods on credit and thus the customers were in fact depending on their benevolence.

It's not necessarily wrong though. Giving credit can still be a self-interest motivated action, in fact it would most often only be that - generating a higher reward to the risk taken by giving it - as banks try to manage. So yes, banks aren't run by benevolence, and they aren't the most benevolent entities on the planet (far from it) - they're just economically rational actors, that are benefited by offering credit.

Interestingly actually, Bitcoin is one of the few asset classes that is de-coupled from its supply/demand.

Unlike say, gold and oil - if that increases in demand, more mines are opened, more sites are evaluated looking for it etc. Whereas bitcoin, the higher (or lower) demand doesn't affect the fact that on average a block is only mined once every 10mins adding 6.25 BTC to the supply... And that throwing more resources into mining it just increases your odds of getting that block reward, but not increasing supply as would happen with the above commodities, or in the example - cheeseburgers.

--To the point - when we have increasing amounts of something like dollars, there certainly are things that can't and wont 'rise with the tide' of those dollars increasing in supply.

Yes, signal would manage that. This is still more privacy preserving than a phone number because the the public key isn't attached to your identity in the way a phone number is. You'd still need to getAllContactsPubKey()->encr(PersonAKey)->SendToAllContacts

So signal knows you requested all your contact pubkeys, that you sent a duplicate broadcast to all contacts, obscuring who it was intended for... that could be 0-n of m persons.

How about no ph number and only UUIDs- You send a msg that only personA can decrypt - but you broadcast the encrypted message to all contacts /+ random UUIDs ... So signal doesn't know who it was intended for exactly, and only personA gets the information. Of course you'd want the app to only alert users getting the broadcast once a received comm is successfully decrypted, otherwise discard.

It's clearly a lot higher data overhead, but that'd acheive phone numberless accounts without signal knowing 100% A is talking to B.... Only that 'A' might be talking to 'B'..Or C..or D.. Or sending decoy msgs intended for nobody.

Settling for phone numbers on a privacy based messenger because it's too hard to do an alternative implem is a cop out I feel. What do you think of the above proposal?

The terminology 'viral gene' is not a made up fear mongering term... It's a plainly descriptive term. It is an encoding for a protein of a virus, in this case the modified c19 spike protein.

Also with regards to fundamental changes to the immune system - this is also true.

""" See: https://en.wikipedia.org/wiki/Original_antigenic_sin

Original antigenic sin, also known as antigenic imprinting or the Hoskins effect,[1] refers to the propensity of the body's immune system to preferentially utilize immunological memory based on a previous infection when a second slightly different version of that foreign pathogen (e.g. a virus or bacterium) is encountered. This leaves the immune system "trapped" by the first response it has made to each antigen, and unable to mount potentially more effective responses during subsequent infections """

Which begs the question of the long term utility and effect of symptom suppressing vaccines (that don't prevent transmission).

If a virus that would make make you very sick is contracted by someone who's vaccinated - They will now be in the 2nd category you describe ~'Sick, but not sick enough to stop them going out' -- They're now giving mobility to that virus that it would have otherwise not had (If they weren't vaccinated and were feeling too sick to go out and about)

So by usage of vaccines that only prevent severe symptoms, we pervert that normal negative feedback loop of virus spread and instead set the parameters to be more likely to propagate severe virus' that would usually keep you sick at home.

Yeah, very interesting... Like you could do a preliminary training that uses really good quality labeled data. Then almost infinitely continue training on a negative only dataset of random noise, which would theoretically hammer out any false positives that can be generated by the model.

Ah I meant, if I drag [n] images at once, eg. cat.jpg, dog.jpg and fish.jpg -- Three fish.jpgs are inserted (Chrome 96). Cheers for the reply, and again, the work looks great :)

Any chance this will become open source? Would love to take a deeper look into it. Looks great though! And something that can use animated webp and do conversions! +BugReport - Dragging multiple webps to add, just adds eg. 5 of the first img in dif rows. (Not tested other formats)

Suggestion: Layers and pasting images from clipboard or uploaded ones would make this a lot more useful too.

That environment of evolutionary pressure is exactly what is being created. An environment where strains with higher fitness for immune escape are more able to spread, in a compounding manner. Particularly with leaky vaccines as we're using for c19 that do not prevent contraction with high efficacy.

Please read this - even just the 'Prevention' section for a real world example. https://en.wikipedia.org/wiki/Marek%27s_disease#Prevention

or read how antibiotic resistance comes about, which seems to be less contentious / politicly affected... But the same evolutionary mechanism is at work.