与「count_1_2」相关的搜索结果

[ #GOT7NOW# ] 유겸의 #RunningTOOgether# 챌린지 💙 #GOT7# #갓세븐# @GOT7official #Yugyeom# #유겸# #TOO# #티오오# #하나둘세고# #count_1_2# #RunningTOOgether# @too_offcl

[ #GOT7NOW# ] 유겸의 #RunningTOOgether# 챌린지 💚 #GOT7# #갓세븐# @GOT7official #Yugyeom# #유겸# #TOO# #티오오# #하나둘세고# #count_1_2# #RunningTOOgether# @too_offcl

Robotaxi Tracker has only 89 riders reporting their rides in Austin, and they’ve logged just 8,152 miles in total. Meanwhile, Austin Robotaxi has probably accumulated 1–2 million miles, meaning only about ~0.5% of the miles have been reported. From that ~0.5% of reported miles, 19 unique unsupervised Robotaxis have been discovered in Austin, and people are laughing about it. Well, to be honest, I’m laughing at you. Imagine I deploy 10 people to take Uber every day and count how many unique cars they discover. On the first day, they report discovering 10 cars. On the second day, they report another 10 cars (20 in total) Several days later, they report that 100 unique cars have been discovered in total (a pretty small number, right?). Then I can finally make fun of Uber: “Look! Uber only has 100 cars. How pathetic!”

Let me lay out the unpleasant arithmetic of the replacement rate, and why a modern society finds it so hard to reach. A population of 100 women in an advanced economy needs 210 children to replace itself. Why? Absent sex-selective practices, roughly 105 boys are born for every 100 girls. Evolution overshoots male births because boys are more prone to early death from accidents and disease. Therefore, of 210 children, about 108 are boys and 102 are girls. Not all girls reach the midpoint of their fertile age: accidents, suicide, homicide, and illness take some. In an advanced economy, about 98% of them survive, leaving 100 women to replace the original 100. Now consider the distribution of children per woman. Imagine 15 women have no children. Five do so by choice, for various reasons (professional, affective, religious). Ten face unfixable fertility problems, theirs or their partner’s. The 10% figure is conservative: the medical literature points to around 13%, and that does not even count male fertility problems. Of the remaining 85, 10 have one child, 60 have two, 10 have three, and 5 have four. I am stopping at four to keep the post concise; very few women in younger cohorts have five or more children, but I could adapt the example to account for them. Hence, the 100 women in this population have 180 children, for a completed fertility rate of 1.8. Interestingly, this is roughly the rate we saw in many advanced economies until the early 1990s, and in the U.S. until around 2008. But we are still 30 children short of replacement! Voluntary childlessness is only 5%. Three-quarters of women have two or more children. Look around: most of your friends will have two, plenty will have three or four. And yet, we are well below replacement. You would not look at this population and call it selfish (is having two kids hedonistic?) or accuse it of losing family values (only 5% of women are choosing voluntarily not to have children). The point is simpler. To reach 210 births, you need a substantial share of women to have three or more children. Two as the “normal” pattern will not get you there. And modern society makes three or more a costly proposition for most families. Of course, current fertility rates in most advanced economies are well below 1.8. But my point is that, under present social arrangements, we should not expect 2.1, even if (to humor last weekend’s debate) we banned smartphones and TikTok. We need many, many more families with three or four children. More pointedly, there is no self-regulating mechanism that pushes a society back to 2.1. The market-clearing analogy many economists use is flawed; scarcity feedback does not work the same way. (Another post on this another day.) And, as I often read, the claim that “nature” somehow regulates current overpopulation is just childish mumbo jumbo. So yes, the arithmetic of replacement rate is unpleasant.

I was given early access to Grok 3 earlier today, making me I think one of the first few who could run a quick vibe check. Thinking ✅ First, Grok 3 clearly has an around state of the art thinking model ("Think" button) and did great out of the box on my Settler's of Catan question: "Create a board game webpage showing a hex grid, just like in the game Settlers of Catan. Each hex grid is numbered from 1..N, where N is the total number of hex tiles. Make it generic, so one can change the number of "rings" using a slider. For example in Catan the radius is 3 hexes. Single html page please." Few models get this right reliably. The top OpenAI thinking models (e.g. o1-pro, at $200/month) get it too, but all of DeepSeek-R1, Gemini 2.0 Flash Thinking, and Claude do not. ❌ It did not solve my "Emoji mystery" question where I give a smiling face with an attached message hidden inside Unicode variation selectors, even when I give a strong hint on how to decode it in the form of Rust code. The most progress I've seen is from DeepSeek-R1 which once partially decoded the message. ❓ It solved a few tic tac toe boards I gave it with a pretty nice/clean chain of thought (many SOTA models often fail these!). So I upped the difficulty and asked it to generate 3 "tricky" tic tac toe boards, which it failed on (generating nonsense boards / text), but then so did o1 pro. ✅ I uploaded GPT-2 paper. I asked a bunch of simple lookup questions, all worked great. Then asked to estimate the number of training flops it took to train GPT-2, with no searching. This is tricky because the number of tokens is not spelled out so it has to be partially estimated and partially calculated, stressing all of lookup, knowledge, and math. One example is 40GB of text ~= 40B characters ~= 40B bytes (assume ASCII) ~= 10B tokens (assume ~4 bytes/tok), at ~10 epochs ~= 100B token training run, at 1.5B params and with 2+4=6 flops/param/token, this is 100e9 X 1.5e9 X 6 ~= 1e21 FLOPs. Both Grok 3 and 4o fail this task, but Grok 3 with Thinking solves it great, while o1 pro (GPT thinking model) fails. I like that the model *will* attempt to solve the Riemann hypothesis when asked to, similar to DeepSeek-R1 but unlike many other models that give up instantly (o1-pro, Claude, Gemini 2.0 Flash Thinking) and simply say that it is a great unsolved problem. I had to stop it eventually because I felt a bit bad for it, but it showed courage and who knows, maybe one day... The impression overall I got here is that this is somewhere around o1-pro capability, and ahead of DeepSeek-R1, though of course we need actual, real evaluations to look at. DeepSearch Very neat offering that seems to combine something along the lines of what OpenAI / Perplexity call "Deep Research", together with thinking. Except instead of "Deep Research" it is "Deep Search" (sigh). Can produce high quality responses to various researchy / lookupy questions you could imagine have answers in article on the internet, e.g. a few I tried, which I stole from my recent search history on Perplexity, along with how it went: - ✅ "What's up with the upcoming Apple Launch? Any rumors?" - ✅ "Why is Palantir stock surging recently?" - ✅ "White Lotus 3 where was it filmed and is it the same team as Seasons 1 and 2?" - ✅ "What toothpaste does Bryan Johnson use?" - ❌ "Singles Inferno Season 4 cast where are they now?" - ❌ "What speech to text program has Simon Willison mentioned he's using?" ❌ I did find some sharp edges here. E.g. the model doesn't seem to like to reference X as a source by default, though you can explicitly ask it to. A few times I caught it hallucinating URLs that don't exist. A few times it said factual things that I think are incorrect and it didn't provide a citation for it (it probably doesn't exist). E.g. it told me that "Kim Jeong-su is still dating Kim Min-seol" of Singles Inferno Season 4, which surely is totally off, right? And when I asked it to create a report on the major LLM labs and their amount of total funding and estimate of employee count, it listed 12 major labs but not itself (xAI). The impression I get of DeepSearch is that it's approximately around Perplexity DeepResearch offering (which is great!), but not yet at the level of OpenAI's recently released "Deep Research", which still feels more thorough and reliable (though still nowhere perfect, e.g. it, too, quite incorrectly excludes xAI as a "major LLM labs" when I tried with it...). Random LLM "gotcha"s I tried a few more fun / random LLM gotcha queries I like to try now and then. Gotchas are queries that specifically on the easy side for humans but on the hard side for LLMs, so I was curious which of them Grok 3 makes progress on. ✅ Grok 3 knows there are 3 "r" in "strawberry", but then it also told me there are only 3 "L" in LOLLAPALOOZA. Turning on Thinking solves this. ✅ Grok 3 told me 9.11 > 9.9. (common with other LLMs too), but again, turning on Thinking solves it. ✅ Few simple puzzles worked ok even without thinking, e.g. *"Sally (a girl) has 3 brothers. Each brother has 2 sisters. How many sisters does Sally have?"*. E.g. GPT4o says 2 (incorrectly). ❌ Sadly the model's sense of humor does not appear to be obviously improved. This is a common LLM issue with humor capability and general mode collapse, famously, e.g. 90% of 1,008 outputs asking ChatGPT for joke were repetitions of the same 25 jokes​. Even when prompted in more detail away from simple pun territory (e.g. give me a standup), I'm not sure that it is state of the art humor. Example generated joke: "*Why did the chicken join a band? Because it had the drumsticks and wanted to be a cluck-star!*". In quick testing, thinking did not help, possibly it made it a bit worse. ❌ Model still appears to be just a bit too overly sensitive to "complex ethical issues", e.g. generated a 1 page essay basically refusing to answer whether it might be ethically justifiable to misgender someone if it meant saving 1 million people from dying. ❌ Simon Willison's "*Generate an SVG of a pelican riding a bicycle*". It stresses the LLMs ability to lay out many elements on a 2D grid, which is very difficult because the LLMs can't "see" like people do, so it's arranging things in the dark, in text. Marking as fail because these pelicans are qutie good but, but still a bit broken (see image and comparisons). Claude's are best, but imo I suspect they specifically targeted SVG capability during training. Summary. As far as a quick vibe check over ~2 hours this morning, Grok 3 + Thinking feels somewhere around the state of the art territory of OpenAI's strongest models (o1-pro, $200/month), and slightly better than DeepSeek-R1 and Gemini 2.0 Flash Thinking. Which is quite incredible considering that the team started from scratch ~1 year ago, this timescale to state of the art territory is unprecedented. Do also keep in mind the caveats - the models are stochastic and may give slightly different answers each time, and it is very early, so we'll have to wait for a lot more evaluations over a period of the next few days/weeks. The early LM arena results look quite encouraging indeed. For now, big congrats to the xAI team, they clearly have huge velocity and momentum and I am excited to add Grok 3 to my "LLM council" and hear what it thinks going forward.

What will you build with CodeBuddy? 🤖 We’re rewarding code builders, developers, and problem-solvers. Join the challenge and tell us what you want to achieve with CodeBuddy. How to enter: 1. Follow this account 2. Retweet 3. Reply in ≤40 words: what do you want to achieve with CodeBuddy? 4. Include hashtag BuildWithCodeBuddy and hashtag TencentCloud. These hashtags will not count towards the word limit above. The most creative entries will receive: - Grand Prize (5 winners): 2,000 CodeBuddy Credits + US$100 Amazon Voucher - Merit Prize (20 winners): 500 CodeBuddy Credits Deadline: June 1, 2026, 23:59 (UTC) Results will be announced via a new tweet on this account on or before June 3, 2026 Eligibility: Open to users in Indonesia, Japan, Singapore, and Thailand only Download CodeBuddy and view full terms via the links in the comments section.