typon

Sound analysis, pretty much agree with you.

If not for Musks genius magic, why do you think Tesla deserves to be priced 7x higher than Toyota? I can see an argument for 2x maybe? But 7x is a massive difference considering Toyota is a not a company run by frauds, it is a serious company with ability to pivot into EV market in due time.

Tesla's current share price is extremely inflated - the market has priced in Elon Musk's genius magic. If this genius magic turns out to be a total fraud, Tesla stock price should go down to levels reflecting the stock price of a similarly sized/revenue producing car company.

To give you an idea:

Tesla revenue: 74.86B

Toyota revenue: 260.13B

Tesla market cap: 498.56B

Toyota market cap: 234.96B

That's roughly a 7.4x difference in expectation vs. reality.

Most people here aren't pissed about the content of the banned account, but the hypocrisy from a self-proclaimed free speech warrior. If he didn't claim to be one, this wouldn't be news or interesting whatsoever.

The best you can do is find a correlated stock (Tesla) and short that instead. I think if Twitter goes bankrupt, you can expect Tesla to enter free-fall.

That's wonderfully framed by Bryan. He's a treasure.

This makes me feel America is back. This is a big achievement and should not be underplayed whatsoever. Whichever country achieves practical fusion is going to be dominant in the next century.

Window/pane/buffer management in vscode is childish compared to vim. Vim motions are incomparibly faster than using a mouse. No, vscode doesn't have everything.

As for 3), we really need to avoid the narrative that these people are idiots or stupid. They are sharks - intentionally malevolent people who have defrauded millions of people. SBF's current defense is that he was too naive and unaware of what was happening at FTX. Obviously its a lie - but we shouldn't help these people spread it.

This is sadly the problem with most new languages.

There were several key insights that have made something like ChatGPT possible relative to traditional neural networks.

* A fixed (but large) vocabulary of sub-word like tokens as inputs.

* Attention mechanism for learning the correlation of words in a fixed sequence window.

* Implementing this attention mechanism in the form of matrix multiplies rather than some other complex math - it allows it to be parallelized and run on GPUs fast.

* Having enough layers of these encoders to have a huge amount of parameter space. ~175B parameters in the case of ChatGPT.

* Feed the model a lot of data - in this case, pretty much the entire internet as text.

* Self supervised learning: we take sentences from the internet and mask out some words and force the network to predict the missing word. Turns out this works extremely well. We don't use the traditional supervised learning inputs -> (prediction, label) paradigm that was the standard 10 years ago.

* RLHF (Reinforcement learning from human feedback). Take generated text from GPT-3 (the underlying generative model) and ask humans to rate different completions. Retrain the model from those ratings.

* A massive compute infrastructure that is able to train this model in a reasonable amount of time, allowing for iteration on hyperparameters. For example, what's the optimal attention head-size? How many encoder layers are good? What should the sequence length be? What should the embedding dimension be? etc. In OpenAI's case, they used thousands of GPUs and thousands of CPUs provided by Microsoft/Azure.

In summary, relatively simple model, parallelizable on GPUs, trained on a lot of data.

I use neovim heavily and I spend the first five minutes of my morning thanking the developers for their hard work - without their selflessness we wouldn't have the best text editor in the world. I was just commenting on whether the purported benefits of using Lua over Vimscript as the core scripting language really panned out. Clearly a lot of people think they did.

I dont know if its me but I find it utterly incomprehensible to read any Lua plugin in neovim. The API is magic and Lua itself isnt a very good language. That is to say, it didnt really get any better than vimscript. Ive written a couple of plugins in vimscript back in the day so i know the pain.

I read a lot on Twitter about the so-called "culture" of Google that prevents them from making AI based products, meanwhile OpenAI has made ChatGPT and is going to replace Google search within the next two months or so. I think this is the same narrative that's being expressed by the GP comment.

To add to your comment, Google has been using BERT to power Search since 2019: https://blog.google/products/search/search-language-understa...

I'm going to guess that the only reason they don't use larger models is because of the compute cost. ChatGPT at 4 billion users with today's hardware is an unsustainable business. However, that thought leads me to imagining: if Google offered Search "Premium" using the latest LLMs, how much would people pay for it?

The difference is Oracle vs. No Oracle.

Videos like this really highlight to me how far ahead Asimo by Honda was. It was released in 2000 before any of the deep learning even existed.

Silicon Valley VC culture still believes in this mythology whole-heartedly - but they add an extra layer of not really understanding what all this actually means because its written in books and those are not blog posts so it's hard to read them

Google's customers: You guys are getting paid?

This has almost become a meme at this point. Anyone in the office who has a problem with their internet on the Macbook, I just send them this:

sudo dscacheutil -flushcache;sudo killall -HUP mDNSResponder

It hurts because all the companies attacking Apple are just the best examples of "The worst guy you know just made a great point". Epic Games, Spotify, Elon/Twitter 2.0 have their own exploitative practices that I can't help but revel in some schadenfreude.

There is no book on data oriented design as far as I know. It would be great if someone could take Mike Actons talk and similar talks and condense the ideas into a book filled with real world examples.

Wikipedia beats megacorp giants like Espn at displaying sports related information in so many metrics and it's not even close. Really a great example of perverted incentives. Espn is trying to show ads, while Wikipedia is trying to show information.

"virtually the entire container ecosystem"

An entire area of software that is completely unnecessary, but because the gods at Google gifted it to us, people have to pretend it's amazing

Design by committee vs. design by singular vision is an age old debate. We have plenty of the former, why not give the latter a try? There are plenty of good languages out there already to use, why does this language have to aim for wide adoption?

* If you want 50ms latency on every interaction with the editor * If you want to send telemetry to Microsoft * If you want slow fuzzy search for massive codebases

Plenty of reasons

Spot on. Absolutely hate this attitude that software sitting there just gathers wear and tear as if it's a mechanical device. Software is written with a particular target platform in mind: x86, ARM, Nvidia GPUs, FPGA soft-processor etc. If the hardware you are running on doesn't change, your software should still function. If the specs of that target platform don't change, your software should still function. If the specs of the target platform change but a hard-working compiler engineer has done the work to make sure your software gracefully uses the new features (for example, a compiler optimizing using AVX instructions), your software should still function.

The fact that most software doesn't continue to function even on the same platform, and on the same hardware, is a massive indictment of the software industry's standard practices.

Zoom's audio + video quality is so much better than Google Meet that I almost feel like companies that use it are getting paid by Google to use it? I have no other explanation.

Film critics, music critics, food critics all need to be directors, artists and chefs?

Check this out: https://github.com/1Marc/modern-todomvc-vanillajs

Pretty good. We have Containers Tabs :)