Amid memory price hikes, is it reliable to buy a decade-old antique computer as your main machine?

Memory prices are skyrocketing beyond belief! The kid was forced to buy an antique computer older than himself—is that reliable? Zhong Jia 10 hours ago It can only serve as a "toy," not a main device.

Affected by rising storage prices, computers are getting more and more expensive. A while ago, a friend of mine planned to buy a gaming laptop, but went back and forth agonizing for a month without making up his mind. The laptop he had his eye on cost just over 6,000 yuan in the first half of the year, and now the price has risen to nearly 10,000 yuan.

On social platforms and digital communities, a large number of users are complaining about current PC prices. Some users with genuine needs, in order to buy computers at low prices, have even started eyeing "antique" computers. Some of these PC devices are even from over a decade ago.

So, with new PC prices staying high today, is buying an antique computer really reliable? The kid bought a second-hand computer older than himself

Recently, we came across a short video on a video platform with over 2 million views, in which a kid born after 2010 takes a budget of a few hundred yuan to a second-hand PC shop to buy a computer. Although the video clearly has staged-for-fun traces, it also provides an answer to this question: with a very low budget, what kind of computer can be assembled from second-hand parts.

Specifically, the core configuration of this computer is: E3 1230 V2 processor, GTX 730 2GB graphics card, 8GB memory.

First, from the processor model alone, you can tell it is not a regular consumer-grade CPU, but a "foreign garbage" part pulled from a server years ago. The E3 1230 V2 is a server CPU launched in 2012, using a 4-core 8-thread architecture with a base clock of 3.3GHz. Its score in CineBench R23 is roughly 750 single-core and 3,500 multi-core. For comparison, the 338H processor widely used in thin-and-light laptops this year scores roughly 2,000 single-core and 16,000 multi-core.

(Image source: Bilibili video screenshot)

Second, graphics performance is crucial, as it determines what games it can play. The GTX 730 2GB is a 2014 product, and its performance is now very outdated; it is basically out of the question for large 3D games, though playing LOL, Golden Spatula and the like at low settings is not much of a problem.

In addition, 8GB of memory is still an acceptable capacity given today's crazy storage price hikes. The reason such a low budget can still fit two 4GB memory sticks is mainly that they are low-spec DDR3 memory, long since discontinued, and second-hand stock is less affected by price increases—especially since they cannot be compatible with motherboards produced in recent years.

Overall, the performance of this second-hand "Frankenstein build" is far behind current mainstream PCs and can only meet scenario requirements with low performance demands. Of course, this post-2010 user's own needs are just playing Golden Spatula, watching movies and so on, and a computer over a decade old can indeed handle that. Antique computers have extremely limited application scenarios

When we pull our gaze back from that post-2010 kid's staged video and carefully study the PC architecture from over a decade ago, we will find that using an antique machine as a main computer is obviously unrealistic. Before giving a conclusion on whether it can be used, it is necessary to first recognize how far it lags behind today's computing needs at the underlying technical level.

First is the hard flaw in instruction sets and multimedia codecs. Processors from over a decade ago (such as Intel's 2nd- and 3rd-generation Core processors) do not even support the AVX2 instruction set, which means many modern productivity applications optimized for newer instruction sets, such as the new versions of the Adobe suite, cannot even run their installers on these old computers; forcing them open will only result in an immediate error and exit.

Even more fatal is the graphics card. Old-generation legendary cards like the GTX 730 or the earlier GTX 750 Ti and GTX 960 could indeed hold their own in the era they were born, but at the hardware level they completely lack support for today's mainstream high-definition video encoding formats. For example, H.265 (HEVC) 10bit, VP9 and even the latest AV1 encoding widely adopted by major streaming platforms today are simply beyond the capabilities of old graphics cards.

When hardware decoding is absent, the task of playing high-bitrate video shifts to the already weak old CPU. The result is that when users watch a 4K or even 1080P 60fps online video, CPU usage maxes out, followed by dropped frames and fans spinning wildly.

Second, the operating system and security ecosystem are no longer compatible either. The newest operating system an antique machine can install is Windows 10. Thanks to excellent compatibility, much hardware from a decade ago can still run Win10. But do not forget that Microsoft's official support for Win10 has basically ended. And to upgrade to Windows 11, the motherboard must support the TPM 2.0 security module, and there is a strict processor whitelist restriction. The vast majority of PCs made before 2017 have already lost eligibility to upgrade to the latest operating system through official channels.

Theoretically, we can forcibly bypass the restrictions by modifying the image, but this may also bring problems such as driver blue screens, UI stuttering and update failures. To put it plainly, this kind of hacky operation is not suitable for novice users.

(Image source: Xiaohongshu)

So what exactly can these cheap antique machines still do? In Leitech's (ID: leitech) view, they can only shine in limited scenarios.

For example, light word processing and web browsing. As long as you install a SATA SSD in the old machine and push the memory to 8GB or even 16GB, it can still barely handle the basic Office trio, WeChat chats, and sending and receiving emails.

There is also running low-load nostalgic games. In addition to the previously mentioned Battle of Golden Spatula and LOL, for classic old games like Half-Life 2 and Left 4 Dead 2, old computers can still run at relatively smooth frame rates at 1080P with medium-low graphics. But old graphics cards don't even support the DX12 Ultimate interface, so playing new games from recent years is basically impossible.

In addition, antique computers are suitable for enthusiasts and geek users to tinker with home servers and NAS. Many users with hands-on skills choose to install Linux on them and put them in a corner to make the most of their remaining value. For example, deploying network drive aggregation and mounting tools like AListServer to turn an old computer into a home media resource center; or using it to run a soft router or act as a lightweight offline download machine. This kind of tinkering can squeeze out the remaining value of antique hardware, but the threshold for tinkering is relatively high, and stability is also rather average.

Antique computers, in many cases, can only manage to run. In terms of stability, energy consumption and user experience, there is no guarantee at all. The ultra-high power consumption of foreign e-waste components, as well as the hidden problems caused by aging motherboard capacitors, often make users' later hidden costs for repairs and replacement parts exceed the money they saved at the beginning. As storage prices continue to rise, will new forms of PCs emerge?

Since antique computers are becoming less and less reliable, why do people still buy them? In the final analysis, it is still because the current prices of new PCs are really a bit exaggerated. Users planning to build a PC or buy a laptop recently will inevitably feel that the current market is hard to describe. Over the past six months or so, key PC components, especially memory and solid-state drives, have experienced drastic price fluctuations.

At the same time, the full penetration of AI Agents in the PC field has also had a lot of impact on the PC market. In our view, AI has not only brought new application scenarios, but also raised the hardware entry threshold for PCs. Now, with the implementation of on-device AI and the popularization of Agents, higher requirements have been placed on PC AI computing power.

More critically, on-device AI has a high demand for memory bandwidth and capacity. For cutting-edge technology enthusiasts, if they want to locally download and deploy open-source large language models with billions of parameters through platforms such as Hugging Face, or run local AI drawing, 16GB of memory is now only considered the passing line, and 32GB or even 64GB is considered relatively ample. Against this background, if traditional low-end affordable PCs are forcibly connected to the on-device AI ecosystem, they will only face the awkward situation of memory being maxed out and background apps being frequently killed.

So, under the impact of soaring hardware costs and surging AI computing power demand, what new forms and product positioning will affordable PCs evolve into in the future? Barring surprises, we believe they will differentiate into these two main types.

The first is highly integrated devices, represented by mini PCs. This type of device integrates components such as the CPU, GPU, memory and flash storage onto a small motherboard. For ordinary individual or family users who do not pursue extreme gaming performance, this kind of device can not only solve desktop space, heat and noise problems, but also meet the needs of daily office work, light editing and entry-level online games with higher cost-effectiveness.

(Image source: photographed by Leitech)

The second is a product type combining device and cloud. Since running AI on-device is expensive and has high memory requirements, future affordable devices are likely to degenerate into a more simplified form, just as many heavy AI users today have already become accustomed to relying on large models in the cloud. As long as there is a screen, a keyboard and a network connection, top-tier cloud computing power can be invoked. Future affordable PCs may integrate cloud AI Agents at the operating system level, packaging inference computing tasks and handing them to cloud servers. This form is expected to pull affordable PCs back to the thousand-yuan price range.

Overall, in this era of rising storage prices and surging AI computing power demand, wanting to assemble an all-around PC on a low budget with the mentality of scavenging foreign e-waste has become increasingly unrealistic. Antique computers from more than a decade ago are destined to become only a compromise choice for geeks to tinker with or for specific niche scenarios. They cannot meet current mainstream needs, let alone reach the basic AI threshold. Future affordable PCs will either move toward highly integrated mini PCs or toward the form of cloud computers. For most people, they are better solutions than antique computers.

Image source of this article: 123RF licensed image library. Infringement report Return Share Related articles All comments