#Computerphile

[fracturedgodhead]

Regular blogging wooooo

We were struggling a lot to read a textbook, for a lot of reasons, but part of it was the fact that we only have a digital copy. We knew that there were tts services out there, and figured that would be easier than reading it with our eyes, so we looked up "pdf text to speech linux" and it turned out to be so easy (as long as you're willing to listen to lower quality output, but some of us feel a certain kinship with the dinky thing, so it's not too bad).

There are 2 programs we ended up needing and they were a set of cli tools called "poppler_utils" and a pretty old program called "espeak"

With poppler_utils we can convert sections of a pdf to raw, somewhat formatted .txt file, and with espeak you can pass in a .txt file and get it to output a .wav, or if you don't tell it to output, it'll just start reading it from the command line. The .wav is more convenient for us though because then we can pause and skip around, and the whole chapter ended up being almost 4 hours, and we can't focus on basically anything for that long without a break.

It's so fun to do I think. It's fun to see the processes run and see the files being generated. God we just love computers so much.

[forgottenbones]

Computing with Soft Robots - Computerphile

[ecrivainsolitaire]

Here's an actual expert going deeper into detail. Please follow creators like him who talk about computers in a rational way, without marketing buzz or pseudoleftist outrage. Being well informed is good for the soul.

A summary of the Chinese AI situation, for the uninitiated.

These are scores on different tests that are designed to see how accurate a Large Language Model is in different areas of knowledge. As you know, OpenAI is partners with Microsoft, so these are the scores for ChatGPT and Copilot. DeepSeek is the Chinese model that got released a week ago. The rest are open source models, which means everyone is free to use them as they please, including the average Tumblr user. You can run them from the servers of the companies that made them for a subscription, or you can download them to install locally on your own computer. However, the computer requirements so far are so high that only a few people currently have the machines at home required to run it.

Yes, this is why AI uses so much electricity. As with any technology, the early models are highly inefficient. Think how a Ford T needed a long chimney to get rid of a ton of black smoke, which was unused petrol. Over the next hundred years combustion engines have become much more efficient, but they still waste a lot of energy, which is why we need to move towards renewable electricity and sustainable battery technology. But that's a topic for another day.

As you can see from the scores, are around the same accuracy. These tests are in constant evolution as well: as soon as they start becoming obsolete, new ones are released to adjust for a more complicated benchmark. The new models are trained using different machine learning techniques, and in theory, the goal is to make them faster and more efficient so they can operate with less power, much like modern cars use way less energy and produce far less pollution than the Ford T.

However, computing power requirements kept scaling up, so you're either tied to the subscription or forced to pay for a latest gen PC, which is why NVIDIA, AMD, Intel and all the other chip companies were investing hard on much more powerful GPUs and NPUs. For now all we need to know about those is that they're expensive, use a lot of electricity, and are required to operate the bots at superhuman speed (literally, all those clickbait posts about how AI was secretly 150 Indian men in a trenchcoat were nonsense).

Because the chip companies have been working hard on making big, bulky, powerful chips with massive fans that are up to the task, their stock value was skyrocketing, and because of that, everyone started to use AI as a marketing trend. See, marketing people are not smart, and they don't understand computers. Furthermore, marketing people think you're stupid, and because of their biased frame of reference, they think you're two snores short of brain-dead. The entire point of their existence is to turn tall tales into capital. So they don't know or care about what AI is or what it's useful for. They just saw Number Go Up for the AI companies and decided "AI is a magic cow we can milk forever". Sometimes it's not even AI, they just use old software and rebrand it, much like convection ovens became air fryers.

Well, now we're up to date. So what did DepSeek release that did a 9/11 on NVIDIA stock prices and popped the AI bubble?

Oh, I would not want to be an OpenAI investor right now either. A token is basically one Unicode character (it's more complicated than that but you can google that on your own time). That cost means you could input the entire works of Stephen King for under a dollar. Yes, including electricity costs. DeepSeek has jumped from a Ford T to a Subaru in terms of pollution and water use.

The issue here is not only input cost, though; all that data needs to be available live, in the RAM; this is why you need powerful, expensive chips in order to-

Holy shit.

I'm not going to detail all the numbers but I'm going to focus on the chip required: an RTX 3090. This is a gaming GPU that came out as the top of the line, the stuff South Korean LoL players buy…

Or they did, in September 2020. We're currently two generations ahead, on the RTX 5090.

What this is telling all those people who just sold their high-end gaming rig to be able to afford a machine that can run the latest ChatGPT locally, is that the person who bought it from them can run something basically just as powerful on their old one.

Which means that all those GPUs and NPUs that are being made, and all those deals Microsoft signed to have control of the AI market, have just lost a lot of their pulling power.

Well, I mean, the ChatGPT subscription is 20 bucks a month, surely the Chinese are charging a fortune for-

Oh. So it's free for everyone and you can use it or modify it however you want, no subscription, no unpayable electric bill, no handing Microsoft all of your private data, you can just run it on a relatively inexpensive PC. You could probably even run it on a phone in a couple years.

Oh, if only China had massive phone manufacturers that have a foot in the market everywhere except the US because the president had a tantrum eight years ago.

So… yeah, China just destabilised the global economy with a torrent file.

[aflo]

yeah, numberphile kinda works as a youtube channel name since they really like numbers. it's a little creepy, but it still scans as something a math major would come up with. computerphile does not. there are already plenty of philes on the computer, we don't need one more.

[thatimageoftomscott]

been a long time tom scott fan but fell off because of uni (funnily enough it's computerphile university... he could be around any corner...) so i didn't know he was on jetlag until i found your blog - thank you so much for putting me on. it's peak television. this is all i'm going to watch for the foreseeable future. i'm overjoyed!

yayyy yippeeee :] super happy to hear that !!!! have fun watching jet lag anon, it really is some prime unhinged tom scott content. sorry uni took tom away from you but considering that he is still very much on sabbatical, its understandable.

[werebutch]

My dad called himself a computerphile 😭😭

[myonlypen]

You know that Tom Scott (or maybe Computerphile) video where he explains how frustrating it is for programmers to implement time zones?

I expect in the Doctor Who universe there's a video like that used internally at UNIT to explain how hard it is to track time lords' regenerations. They have to change their software every time they learn that their genders can change or they can regenerate into two people, or can regenerate into previous incarnations, or can grow a new version of themselves from a hand, or can have more than 13 regenerations.

[penipso]

Computerphile is the funniest channel it's like let me get a completely off the cuff explanation of a computer science concept by some tenured British professor and watch him and the video editing team struggle to make the concept digestible to the layman

[mirthalia]

That's actually one of the things DeepSeek has been kind of revolutionary about! So of course, there's a Computerphile vid about that too:

nothing funnier to me than when AI does math wrong. like I get why it happens, it's a language model that's treating the numbers you feed it as words rather than integers and then giving you an answer based on how those words typically appear in a block of text instead of actually performing a calculation. but the one thing computers are genuinely incredible at. you fucked up a perfectly good calculator is what you did, look at it it's got hallucinations

[e5e5e5e5e5]

Ai Will Try to Cheat & Escape (aka Rob Miles was Right!) - Computerphile

OH WOW FANCY THAT

[skyshoes]

DeepSeek is a Game Changer for AI - Computerphile

[enetarch-programming]

Creating Your Own Programming Language - Computerphile

[david-goldrock]

Computerphile made a great video about how it works

There is something deliciously funny about AI getting replaced by AI.

tl;dr: China yeeted a cheaper, faster, less environmental impact making, open source LLM model onto the market and US AI companies lost nearly 600 billions in value since yesterday.

Silicone Valley is having a meltdown.

And ChatGTP just lost its job to AI~.

[levysoft]

Scambio di chiavi segrete (Diffie-Hellman) - Computerphile

Come si scambia una chiave segreta in chiaro?

Cos'è la crittografia a chiave pubblica e come funziona realmente. Se vuoi una spiegazione reale, buona e visiva di come funziona, con semplici calcoli matematici, guarda il video di Computerphile su Diffie-Hellman: https://www.youtube.com/watch?v=NmM9HA2MQGI

(via Secret Key Exchange (Diffie-Hellman) - Computerphile - YouTube)

[skipeon]

Has Generative AI Already Peaked? - Computerphile

Has Generative AI Already Peaked? – Computerphile Summary The discussion centers on the limitations of generative AI, challenging the notion that simply adding more data will lead to significant advancements in AI capabilities. Highlights 📊 The idea of general intelligence through more data is debated. 📉 A recent paper suggests data requirements for performance may be astronomically high. 🐱…