TheGeminon

In the paper DeepSeek just says they have ~800k responses that they used for the cold start data on R1, and are very vague about how they got it:

> To collect such data, we have explored several approaches: using few-shot prompting with a long CoT as an example, directly prompting models to generate detailed answers with reflection and verification, gathering DeepSeek-R1-Zero outputs in a readable format, and refining the results through post-processing by human annotators.

Maybe they’ll try to build up traffic to your site from those domains and then push to sell them to you/extort by removing the redirects?

It is also unavailable from Canada.

If you read a bit further he excludes instances like that and listed films with only a single (likely intentional) title drop.

> Okay, so those are the problems. What’s the solution?

> If you need to perform a case mapping on a string, you can use LCMap­String­Ex with LCMAP_LOWERCASE or LCMAP_UPPERCASE, possibly with other flags like LCMAP_LINGUISTIC_CASING. If you use the International Components for Unicode (ICU) library, you can use u_strToUpper and u_strToLower.

Devaluing the new currency by adding lesser metals will also devalue existing currency that is "pure" as you aren't able to trust the value of the currency anymore, so the value of the existing pool of money will drop.

Its at a smaller scale, but it can be seen with counterfeit currency today. Cash-heavy businesses have to absorb whatever amount of counterfeits they accept, so they are really valuing your dollar at $0.99 if they might have to throw it out.

There are ways to gauge the confidence of the LLM (token probabilities over the response, generating multiple outputs and checking consistency), but yeah that’s outside the LLM itself. You could feed the info back to the LLM as a status/message I suppose

Democracies are more than capable of passing laws that violate the rights of minorities.

So this would be a worthwhile case if a youthful physicist was bringing the complaint?

As per the article, they are arguing that seniors are particularly affected by climate change due to heat waves, and these are causing deaths.

The 2022 heatwaves had hundreds of deaths attributed to climate change in Switzerland: https://lenews.ch/2023/07/08/climate-change-behind-60-percen...

With 37,000 Palestinians marked as suspected militants, it would mean they expected up to 555,000 - 740,000 civilian casualties.

This is also reported on by The Guardian https://www.theguardian.com/world/2024/apr/03/israel-gaza-ai...

According to Wiktionary in English that didn’t appear until the 14th century (although that is earlier than the article suggests)

https://en.wiktionary.org/wiki/heart

The ideal way is to release a security advisory that some CVE patches are coming with no details, so that upgrades can be prepped coordinated for a known patch day.

You just need to focus niche and upmarket, OpenAI is e.g. never going to make that "clone your chats and have your LLM-self go on pre-dates" app that went around Twitter.

It doesn’t need to be all or nothing though, humans are social creatures. If rates are similar, having 5-10% of your tribe being more exploratory could be advantageous, even if the rest are picking up slack in other areas.

Unsure if it’s pseudoscience, but similar to the argument around morning vs. evening preference being related to having someone alert for the tribe at the right times.

Outperform is dependent on the RAG approach (and this would be a RAG approach anyways, you can already do this with smaller context sizes). A simplistic one, probably, but dumping in data that you don't need dilutes the useful information, so I would imagine there would be at least _some_ degradation.

But there is also the downside of "tuning" the RAG to return less tokens you will miss extra context that could be useful to the model.

Where are you getting that the company is worth $50m? The article says it’s a “$100 million-a-year” business

From my (limited) reading it seems like this is a UN treaty to allow for better co-operation between states when prosecuting/investigating cybercrime.

The main issues I see is that there is little distinction in what constitutes a crime that would fit under the "cybercrime" banner, potentially allowing a state that has e.g. anti-LGBT laws to request assistance in prosecuting that crime by e.g. getting de-anonymized data from a US-based social network.

That confused be a bit at first too, but that ranking seems to be for how widely they are used (not sure the exact metric), not the ranking of how good they are (which is the dbms of the year page)

There is a more detailed explanation at https://unsloth.ai/introducing

You don’t need to apply the license until after the public release, before then it would just be sharing code with your customers under whatever agreement you want

That kind of decoding is a bit different though. For one, the tokenization process makes encodings difficult to handle (unless it’s trained on a lot of pairs).

This would be more akin to asking ChatGPT to help build a black box parser for base64, not asking it to decode it itself.

The problem is that it’s doing a search of your relatively niche problem, and probably getting pretty poor results. The text from the search is then more highly weighted than the base model, but with relative junk, so it performs better without the additional (unhelpful) context.

You see this with Bing search on ChatGPT as well, and I’ve seen it in my own projects.

This seems like a great way to pull out session identity from cookies and sidestep a lot of the baggage there.

I would caution framing it as a secure replacement for cookies, it’s a secure replacement for session ID token cookies. Tons of cookies aren’t just opaque IDs.

They could also check the physical meter for that, or if they are wardriving for the signal they could likely pick up the presence/lack of WiFi/5G/LTE signals which would probably be even more accurate than power usage.

North American credit cards can have very strong benefits (which merchants pay for in their fees) giving 1-5% back on purchases, and other deals/discounts/perks, so there is a lot more perceived value to using them and (intending) to pay it off in full.

But that also means it’s easy to just keep using the card, and many people don’t actually use the cards to their full extent to get the full perceived value either.

Upvotes are meant to signify productive conversation, not necessarily that you agree with it.

That’s the problem with anything that is defined by the symptoms, rather than the cause. As soon as you find definitive evidence of what causes a subset of SIDS, it’s no longer really SIDS but “improper bed safety” or “undiagnosed genetic condition”.

Ayman al-Zawahiri was killed by a US drone strike

https://en.wikipedia.org/wiki/Killing_of_Ayman_al-Zawahiri

> Instead of bouncing back and forth between the client and server, we do our database query as part of the initial request, sending the fully-populated UI straight to the user.

> But hm, how exactly would we do this?

Couldn't help but chuckle. What mysterious technology could allow us to send an entire fully-populated UI over the web?