Technology

https://archive.vn/UMpLO

https://archive.ph/xk7In

https://archive.ph/fvHq5

https://archive.ph/XyNXM

This is a perfect illustration of why government direction of labor towards productive purposes is essential and why this task cannot be left up to the markets.

The government started to crack down on the quant trading industry amid economic slowdown in an effort to move the economy away from the financial sector and towards manufacturing.

High-Flyer was one of the quant traders that was facing extinction. Instead of giving up, they spun out Deepseek, an AI lab using their existing talent and 10k GPUs they were already using for trading. They didn't even need any VC funding.

DeepSeek also proves you don't need billions in funding, as it's being developed on the cheap compared to commercial models like ChatGPT. The team is able to produce a cutting edge model using a fraction of the resources its competitors need.

DeepSeek approach fundamentally changes the economics of the market and makes OpenAI's strategy obsolete.

Scores of projects are starting to use DeepSeek model which leads to an ecosystem being formed around it, turning it into a standard setter. The model is open and free for anyone to use making it more appealing to both public and private enterprise, and it don't require massive data centers to operate. While large versions of the model still need significant infrastructure, smaller versions can run locally and work well for many use cases.

Another aspect of open source nature is that it amortizes the development effort. The whole global community of researches and engineers can contribute to the development of the model. On the other hand, OpenAI has to pour billions into centralized infrastructure and do all the research to advance their model on their own.

The competition here is between two visions for how AI technology will be developed going forward. DeepSeek's vision is to make AI into an open source commodity that's decentralized and developed cooperatively. OpenAI vision is to build and expensive closed system that they can charge access for.

Traditionally, open source projects that manage to gain significant momentum have always outcompeted closed source software, and I don't see why this scenario will play out any different. This calls into question the whole $500bn investment that the US is doing into the company. The market will favor cheaper open model that DeepSeek is building, and it will advance faster because it has a lot more people contributing to its development.

And looping back to the initial point, DeepSeek would not have existed if the government in China didn't crack down on trading industry which redirected labor towards productive endeavours.

@RnaudBertrand: "The right question to ask after what happened these past few days isn't how Deepseek is going to make money. It's how OpenAI will.

Deepseek isn't the one that needs to make a ROI on half a trillion dollars worth of data centers (or whatever fraction of that amount actually materializes) with a product that's now offered free by the competition 😏

And that's probably exactly the point of Deepseek's strategy: to fundamentally change the economics of the market so as to make OpenAI's model obsolete.

Let's play this out and assume that Deepseek's strategy works out, and from where I stand it's looking like it's starting to. What "working out" means is scores of AI projects now starting to use Deepseek's model (in Open-Source or via their API) to power their AI endeavors, resulting in an ecosystem effect and them becoming a standards setter.

It's also them proving that many AI applications don't require massive data centers. While the most powerful models still need significant infrastructure, Deepseek's smaller versions can run locally on personal computers and gaming PCs, making OpenAI's $500B investment look highly questionable for many use cases.

And there's a brilliant strategic angle here: while OpenAI pours billions into centralized infrastructure, Deepseek is democratizing AI by enabling local deployment. This allows them to expand their reach without massive infrastructure investments: their users make it for them.

If you're OpenAI, this all ought to make you sweat. You're basically IBM in the late 1970s watching personal computers starting to democratize computing. Your $500B bet on centralized computing power might be happening just as the market shifts toward distributed, commodity AI.

And you're stuck: you can either dramatically cut your prices to compete (as a reminder, Deepseek charges just 3% of OpenAI's prices for their API calls, good luck making ROI on $500B of infrastructure if you match them), or try to differentiate by coming up with better models - bearing in mind that Deepseek has a track record of catching up to your models in a matter of days or weeks.

All in all it looks like OpenAI's expensive infrastructure might end up being the ultimate liability rather than the moat they hoped for.

To come back to the original question of monetization, what this all means is that Deepseek's approach is almost like that of guerrilla forces choosing terrain that turns an enemy's superior firepower into a liability. They're changing the game to impose their vision of AI as an open commodity that runs everywhere versus OpenAI's vision of a closed service controlled centrally. While OpenAI builds massive, expensive bases, Deepseek is empowering local resistance through distributed, efficient deployment. History shows how that usually ends."

I wanted to drop an update mainly to record the solution for posterity but also for thanking all those who took the time to respond and help me out because I got some critical pointers that way without which I would have been lost. There are a lot of similar posts on Windows' support website but they feel like they have been written by customer service people who just ask you to reinstall Windows or by illiterates. There is never an attempt to give any reason behind their suggestions.

This post and its crosspost is what I am referring to: https://lemmygrad.ml/post/6734778

Long story short, I have a dual boot Linux and Windows setup. I reinstalled my Linux OS to try out NixOS, went back to Void. But since installing NixOS, neither rEFInd nor GRUB was able to find the Windows installation. This is what the disks' layout looked like throughout the process. nvme1 houses the Linux install and the other disk has Windows.

$ lsblk -f
NAME        FSTYPE     FSVER LABEL [...] MOUNTPOINTS
nvme1n1     zfs_member 5000  zroot [...]
├─nvme1n1p1 vfat       FAT32       [...] /boot/efi
├─nvme1n1p2 swap       1           [...] [SWAP]
└─nvme1n1p3 zfs_member 5000  zroot [...]
nvme0n1
├─nvme0n1p1
└─nvme0n1p2 ntfs

At the time of the original post, I had mistakenly assumed that Windows had its own partition related to booting somewhere. I made this assumption because Windows engages in a lot of proprietary weirdness. This was incorrect. The files pretaining to booting Windows were located in nvme1n1p1 alongside the Linux boot files. In the process of installing Linux, I had reformatted nvme1n1p1 because of which Windows' boot files were lost. The fix for this boiled down to restoring these lost files and running {grub,refind}-install so that a boot entry for Windows was added by the respective bootloader.

Before this, I had another problem at my hands. Somehow I had made nvme1n1 use a MBR partition table. Because of this, I could not follow any guides that used a Windows recovery media to install the boot files. I don't know how that happened because I always use GPT. So I had to reformat the disk and install the operating system again. I should have saved a ZFS snapshot and restored from it to avoid a lot of headache but in my panicked state I forgot to do that and had to reinstall and set up Void Linux from scratch.

Then I followed this guide that I found in a Windows support thread reply: https://woshub.com/how-to-repair-deleted-efi-partition-in-windows-7/

For some reason, Windows did not accept that nvme1n1p1 was a valid EFI partition. So I had to reformat this partition using the diskpart tool that is available in Windows' command prompt. The series of commands looked something like this:

diskpart
list disk
select disk 1
list partition
select partition 1
delete partition
create partition efi size=512
select partition 1
format quick fs=fat32 label="System"
assign letter=P
exit

Deleting and recreating partition 1 may have been unnecessary. But since the command for creating it contained efi in it I didn't want to take chances.

The Windows drive was already mounted at C:\. So I ran this command:

bcdboot c:\windows /s P: /f UEFI

This copied the Windows boot files and made Windows Boot Manager the top entry in the boot order. Now I was able to boot into Windows but not into Linux since rEFInd had been wiped in the process.

Then I had to create a Linux recovery media and reinstall rEFInd. Now rEFInd was able to detect Windows and add an entry for it in the bootloader.

All was good in heaven and on Earth.

Thanks everyone for the help again.