timthelion

In Czechia they are just littered everywhere on the sidewalk.

I wonder if model merging allows for peer to peer collaborative learning ala folding at home. I also wonder if it can have protective effects against catasrophic forgetting similar to the way DNA is merged in sexual reproduction.

Much more importantly for ARM phones video isn't decoded in the CPU at all, the SOC has video codecs built in. This allows things like all day netflix/hbo without draining the battery. AMD cpus don't have video codec coprocessors as far as I know. Not sure about GPUs though...

yeah, you're right, major versions are a pita...

In my experience it is less that 1 minute but may not be zero downtime...

Amazon is a blogspam drop shipping site in Europe

Yeah, here in Czechia I always look at https://www.heureka.cz/ first.

Yes I strongly agree. I remember a friend bought coloidal silver cleansing spray at a pharmacy. She thought it was a normal antiseptic spray and used it to clean her kids wounds. She was very surprised when I explained it was just overpriced distilled water. I'm in Europe so there were no health claims on the packaging but other than that there wasno way to know. It is especially confusing in that case since silver is a real antiseptic, so it's not like obviously non plausible.

While a formalized concept of 'turing completeness' which could prove the equality of sufficiently advanced LLMs will not exist yet, it will become clear that all models of a certain capability are equally capable(just as all CPUs are equally capable) and LLMs will become a comodity traded by speed, price and context window width. Furthermore, it will be realized, that beyond a certain size, larger context windows lead to diminishing output quality and we'll start talking about the theoretical limit for context window size (this will be a function over the amout of training data). OpenAI will still be a valulable company, but there will be at least 5 LLM providers to choose from and none will significantly outshine the rest.

At least one company will raise over $1billion on entirely fraudulent AI (using gpt's API and sleight of hand).

If that forces the US to simplify their tax system it might end up being the only positive that comes from this war.

I don't understand how it could ever make sense to capture dilute CO2 from the atmosphere when there is concentrated co2 being released from power plants. Isn't it always more effective to capture at the source with the exception being non fixed sources like cars?

Cryptographic signatures exist. This is a trivial exploit to solve...

Yes and no. The Czech law is very vague saying that contracts must be written in good will and be understandable by the signatories. It is the Czech supreme court which wrote up a legal test for understandability which you can find here https://www.epravo.cz/top/clanky/absolutni-zakaz-smluvnich-p...

I've translated the test using deepl:

Translation results

"In practice, the principle of fairness is manifested, inter alia, by the fact that the text of a consumer contract, especially if it is a form contract, should be sufficiently legible, clear and logically organised for the average consumer. For example, contractual terms must be of sufficient font size, not be significantly smaller than the surrounding text, and not be set out in sections which give the impression of being irrelevant. This principle of fairness also applies to the application of general terms and conditions. As stated in paragraph 9, general terms and conditions may also be applied in consumer contracts, but such application is subject not only to the formal limitations mentioned but also to restrictions as to content."

Another great on is that text size has legal meaning here. The larger/darker the text the greater the legal weight. So if the contract says two contradictory things, the larger text wins out...

Having grown up in the US, my absolute favorite law in Czechia is the one that says the advertised price has to equal the price on the bill. In the US, you get a $20 cell phone plan and the bill is for $60 after fees. In Czechia the price is always exactly as advertised.

Well thats what the sigularityis all about, eh?

Now I'm making bank. My anual sallary is around what mozilla CEO makes in two workdays.

I did, went from working for a small non profit in the sustainable transport/urbanism sector to management in the telco industry. Sallary jumped almost 10x.

Those two.scenarois dont compare well,because people can still drive just fine at a lower speed...

it's good not to be able to give a neuteral rating, if there is a center lazy raters skew neutral.

Do you get unsollicited phone calls in Europe? In Czechia it almost never happens, maybe once every three months...

* When you filter the leaderboard by models that are exactly 7b parameters

Is the reputational harm any greater than the reputational harm in being openly gay? The 'reputational harm' is that bigots will judge you.

Is this somehow related to kitboga? I don't see any indication that it is...

Who does that money go to?

There are the up down thumbs and automatic sentiment analysis as a test.

Calling that a conspiracy is like saying its a conspiracy theory that Meta shows different people different Ads. I'd be more concerned if OpenAI WASN'T constantly trying to tune their models. Its literaly their job to tune the models.

How does google get the right to create net TLDs? Can I create a TLD?

I think we need a new type of status page or at least a public version number on llms, yesterday for me GPT4 started giving nonsense super generic answers, like it was hardly reading what I wrote, and today it is back to top notch performance. I think they were trying to make the model more efficient or something but I just saw a massive decrease in the quality of output. From my side though, there is no version number except for "4"...

Amazon deleted books from people's kindles though.