anavat

"Capability stays inside the curriculum" implies that even much more advanced models are not able to go far beyond their pre-training data. Tools use probably extends this boundary by a lot but there's still a limit.

Yes, it was a criminal offense since 1961 (I didn't hear about a special law for gypsies but it could be). So there were homeless and unemployed, but only those who really wanted to be. It took effort.

It was far from utopia but it wasn't orwellian. It was different in a way that's hard to explain if you didn't experience it. The lack of public goods, lower living standards, and the oppressiveness of the state were somewhat compensated by the peace of mind coming from knowing that you couldn't become unemployed or homeless. This is what a lot of people miss when they say they miss USSR.

No, it actually continues running headless on the host, and you can reconnect from another laptop or mobile phone, or even ssh to the host and attach to the session. At least Codex desktop app works this way.

Sorry, it was more of a "shower thought" than a serious statement. The borders are blurry. Maybe it is alive. Maybe everything is alive. Maybe it is all relative. Humans classify as separate entities only on certain zoom level - zoom in and you get a symbiotic colony, zoom out and all you see is a population. Maybe AI is alive but only in the short time during inference. Maybe a continuously running claw is closer to a multiple generations of an organism going through evolution on each inference step. I am just musing.

It comes down to "what stays the same when all its atoms are replaced" and "how can we tell it is the same entity and not just an identical but different one". I think the answers are "shape" AND "continuity".

E.g. all "atoms" of a glider in Conway's Life get replaced every couple turns but an observer can tell it is still a glider because it keeps the shape, and it the same glider because it continues it's previous state.

This makes AI not quite alive because it's missing the continuity.

Because of compound efficiency and technological enablement.

Or, risking to beat the metaphor to death, because over a span of time I'll cross many more deserts than I would have on a camel, and because I'll cross deserts that I wouldn't even try crossing on a camel.

It is simple. Continuing your metaphor, I have a choice of getting exactly where I want on a camel in 3 days, or getting to a random location somewhere on the other side of the desert on a helicopter in few hours.

And being a reasonable person I, just like the author, choose the helicopter. That's it, that's the whole problem.

Okay, I'm gonna shoot myself, "ensloped" it is.

"I find your email deeply ensloping."

"This marketing campaign is going to enslope a lot of people."

"Feeling ensloped, I closed Instagram and looked out the window".

There needs to be a word for the feeling of sudden realization that you're reading an AI-generated text (or watching an AI-generated video) where you expected it to be human-authored.

Yeah, strong crypto bubble vibes. Everyone is building tools for tool builders to make it easier to build even more tools. Endless infrastructure all the way down, no real use cases.

You may want to check out https://purelymail.com who are pretty close to your target pricing. I've been using them since forever and have no reasons to switch but it is always great to see more players in the niche!

Holy... That just unlocked a memory. When I was a 10 y.o. kid and a member of Young Pioneers organization in the late Soviet Union, we had a routine elderly care assignment. There was literally a task "to help N lonely elderly people next week"! We were split in pairs and provided with a list of addresses of nearby "targets" and a general guidance on what this help may constitute, like helping them with groceries, walking a dog, house cleaning, etc. And I remember there was a nice old lady who were treating us with tea and pies, and showing us some old photos, and who we then visited few times outside of the assignment just because she welcomed us and we liked it there. For me it was just some secondary quest that I barely registered and completely forgot about until now, but now I wonder if it was a big deal for her

Thanks for working on this! Language learning really needs a breakthrough.

Now, I tried the web app and chose to learn Greek as a beginner. And while I had better experience with your app than with ChatGPT or Gemini voice modes, I still got lost 5 minutes in because the AI tutor doesn't seem to have a plan for me, nor does it "see" my struggles. For example, after asking me about a hobby, it gives me a long sentence in Greek about how how it is nice to hike in mountains. Being absolute noob I cannot reply to it, nor even repeat it. And I don't even know what it is expected from me at the moment. A human tutor here would probably repeat a part of the sentence with a translation and ask me to repeat, or would explain something. The AI just sits there waiting for me to make a sound, and when I make it, it goes on on a tangental subject of beach vacations. :)

Again, this is still relatively not bad, and I'm going to give it another try.

I don't buy it makes me dumber. It just makes me worse at some things I used to do before, while making better at some other things. Often times it doesn't feel like coding anymore, more like if I were training to be a lawyer or something. But that's my bet.

My take is that AI's ability to generate new code will prove so valuable, it will not matter if it is bad at changing existing code. And that the engineers of the distant future (like, two years from now) will not bother to read the generated code, as long as it runs and passes the tests (which will also be AI-generated).

I try to use AI daily, and every month I see how it is able to generate larger and more complex chunks of code from the first shot. It is almost there. We just need to adopt the new paradigm, build the tooling, and embrace the new weird future of software development.

For humanity? It is the same reason most of us don't grow our food or make our shoes. Specialization makes a lot of sense. Western tech wouldn't have been able to grow as fast if electronics manufacturing were not concentrated in China. It is (was?) a win-win.

My guess is this is an artifact of the RLHF part of the training. Answers like "I don't know" or "let me think and let's catch on this next week" are flagged down by human testers, which eventually trains LLM to avoid this path altogether. And it probably makes sense because otherwise "I don't know" would come up way too often even in cases where the LLM is perfectly able to give the answer.

> But that doesn't scale at all.

It doesn't scale if performed by a human. But what if... we employ AI to conduct the voice exams?

Western media has full control over information and if Ukrainians would do something like this, no one would even know.

There were multiple cases of Ukrainians doing something evil or borderline evil that were swept under the rug. One of the recent examples is the cassette munition explosion over a beach in Sevastopol, which killed few children and wounded a hundred of people. I stopped following the war closely but it got my attention because that's my home town. And in this case I even agree that the rocket wasn't specifically targeting the beach (that would be stupid), it was likely targeting the nearby airbase; but that's not the point.The point is that every single time something like this happens, it gets silenced.

There are multiple high-quality videos of the explosion recorded from different angles. On Reddit, a high-quality video of an even like this, surreal and frightening, would otherwise have been upvoted to skies. But not when it puts Ukrainians in bad light! One the next day, as a random Reddit user, you'd never even know about this event (I wonder how many people know about this at all).

And again, this is just one example. I can probably collect few hundreds of cases likes this over the first two years of the war, where as a Western media consumer you would never know about something that could potentially change your opinion on the conflict. And just like this, you're being manipulated. Of course, so are the Russians who solely rely on Russian news sources. The only way to know the truth is to follow both sides closely, especially to what each side hides and silences. You'd be surprised.

Well, it really tells you haven't tried buying drugs on Telegram. It is all scammers, every single one of them (maybe with the exception for prostitutes but I bet most of these are scam too). There is pretty much 0% chance you'll buy drugs or fake documents using the geo search. They will scam you for a transfer and disappear. It really is no that different from a spam email, just different media and targeting.

That said, you CAN buy drugs on Telegram, sure, it is just not as easy as everyone seems to think. You need to know the account name of a service that delivers in your area, you need to be reeealy careful when typing the account name because for every real drug seller account there are multiple fakes with slight variation in the name that fish for people using search, and then even if you have a verified account and they sold you drugs last month, there is like 30% chance that the account have been compromised and now you are talking with scammers again.

"Find me a dealer close by" is a sure way to get scammed. I'm not saying you cannot buy drugs on Telegram, just not via the location search.

Oh, I am not a settler, I was born in Sevastopol and lived here most of my life, as most people I know. It is hard to be sure, but I'd say, only about 10% of the current population came after 2014, the rest lived here before and then took Russian citizenship.

Cancelling property deals after 2014 would effectively mean most will still be able to live in their homes and apartments but will never be able to sell anything. It took me 5 years to change Ukrainian documents to Russian documents (property documents, passports, driver license, everything), and I still have not received and will probably never receive ownership on the small workshop that belonged to my father's company because it requires something from Ukrainian archives, which is pretty much impossible to obtain. I spent months of my life standing in queues in government agencies and in the court. I cannot imagine changing everything back, it will take decades!

And then, what about the kids who were born here after 2014 and have no Ukrainian documents? What about tens of thousands who lived here all their life but who always were Russian citizens, like my father in law? Sevastopol has been a Russian navy base since 1783, there are so many people who are kids and grandkids of retired navy officers, who were born and lived here but never had Ukrainian citizenship. What about their families, and everybody who had some relation with Russian navy? Half of guys my class in 1994 went to the local Russian navy academy and became Russian navy officers, no chance they or their families can re-integrate in Ukraine. And then the Russian military shipyards, and the local branch of Moscow State University and many many other Russian institutions which were functioning here since forever and all the people working for them and their families... Russia has always had so much presence here, I honestly cannot imagine Ukraine taking over Crimea without expelling or imprisoning half of the population.

Sorry for the brain dump, I didn't mean to make a statement or prove anything, it is just such a mess it hurts to even think about this all.

In the unlikely event of retaking Crimea, Ukraine is going to prosecute collaborators, cancel _all_ property deals after 2014, and expel Russian citizens who never had Ukrainian citizenship. I have received Russian citizenship, some of my family members have been Russian citizens all their life, and I have inherited, sold, and bought property after 2014. And while it is not clear who is going to be considered a collaborator, nobody who received Russian citizenship is safe (which means nobody at all).

Well, I don't know what to say really. I have no idea how this all ends, and I fear it is far from over yet. I don't support Ukraine in this conflict for different reasons, one of which is that in case of their victory my family will loose everything and I'll end up in jail (and that's best case). But I cannot support Russia either because I cannot support the idiots who decided to start this war, because I cannot support the way this war was fought, and because if Russia wins I'll be sealed in one country with these idiots forever. And in the case the conflict will get frozen at some point (which is probably how it'll go because I cannot see either side winning on the battlefield), it'll just continue later. In short, I think we're screwed either way.

The water situation is way overblown. Crimea did fine since 2014 when the water was cut. There were shortages in the central and north parts but mostly because of high demand from the local industry (Crimean Titan plant) and agriculture (it is hard to believe but they were planting rice over there). It also helped that due to the climate change we had unusually heavy rains and even floods in the southern parts. Because of that, water supply in Sevastopol hasn't been rationed even once even though the population has increased, although it was a common occurrence before 2014.

Yes! Look up the mystery of the SolidGoldMagikarp word that breaks GPT3 - it turned out to be the nickname of a redditor who was among the leaders on the "counting to infinity" subreddit, which is why his nickname appeared in the test data so often it got its own embeddings token.

What if we loop it to itself? An infinite dialog with itself... An inner voice? And periodically train/fine-tune it on the results of this inner discussion, so that it 'saves' it to long-term memory?