Pearl may have a ready response, but the study takes it on directly. Together AI, which announced an exclusive partnership in May, framed the deal as letting "every GPU cycle powering AI training and inference" also mint the PRL token, and it now offers a discounted Gemma-4-31B-it-pearl inference endpoint subsidized by mining proceeds. Basu counters that this is financial arbitrage rather than useful mining, because Together AI's own GPUs perform that inference separately from the mining network, with PRL revenue used to trim the endpoint's price. The 8,012 mining workers he measured, he says, produced none of that inference themselves.
The study’s conclusion is not that Proof-of-Useful-Work is impossible, but that Pearl’s current design leaves a major enforcement gap. The protocol enables useful work in theory, but it does not require it in practice. That leaves Pearl in an uncomfortable middle ground — it performs real computation, but according to the paper, the network currently has no way to prove that the computation is useful AI work rather than cryptomining with AI-shaped math.
Follow Tom's Hardware on Google News , or add us as a preferred source , to get our latest news, analysis, & reviews in your feeds.
Etiido Uko is a news contributor for Tom's Hardware covering the latest updates in big tech and the PC industry. He is a mechanical engineer and senior technical writer with over nine years of experience in documentation and reporting. He is deeply passionate about all things engineering and technology, and is an expert in gadgets, manufacturing, robotics, automotive, and aerospace. ","collapsible":{"enabled":true,"maxHeight":250,"readMoreText":"Read more","readLessText":"Read less"}}), "https://slice.vanilla.futurecdn.net/13-4-24/js/authorBio.js"); } else { console.error('%c FTE ','background: #9306F9; color: #ffffff','no lazy slice hydration function available'); } Etiido Uko Social Links Navigation News Contributor Etiido Uko is a news contributor for Tom's Hardware covering the latest updates in big tech and the PC industry. He is a mechanical engineer and senior technical writer with over nine years of experience in documentation and reporting. He is deeply passionate about all things engineering and technology, and is an expert in gadgets, manufacturing, robotics, automotive, and aerospace.
Pierce2623 It’s crypto. Of course they’re not doing real useful work. It’s mining, period. Its just not hashing like BTC, which coincidentally is also very bandwidth intensive like inference. In fact the claim that the GPUs are using very little of their available bandwidth is quite out of the ordinary for GPU mining. Reply
JC5000 Ban this scourge on our society. You people are LITERALLY digging up coal so computers can toil away doing NOTHING Reply
JC5000 JC5000 said: Ban this scourge on our society. You people are LITERALLY digging up coal so computers can toil away doing NOTHING It's like a mental illness Reply
hwertz I'm not suggesting, at all, that this is a good way to get anything done. But, people pay to put jobs into Pearl (and then the people running the mining rigs get some pay for doing the computations.) And they do have those inference endpoints so you can use this to run real workloads, i.e. it's not some 'theoretically it could be used for work' but no practical example of doing it. Why would people spend good money putting useless workloads into the system? I mean, if you want to pay a cloud provider directly and run nonsense on CPU or GPU, they don't verify the workload they are running either, there'd just be this assumption that people usually wouldn't pay them to run 100% useless workloads. Again, I haven't looked into Pearl to see if it's REALLY useful for these workloads, or if it's like some of those blockchain style things where the blockchain/crypto bros think it's the schizz but in reality it's like a distributed database that's way more expensive per transaction and like 1/10000th the speed. (i.e., if you use that gemma endpoint, do you get a response in a reasonable length of time? Or does it spend "a while" shuffling things around and you get a response way way later?) Reply
usertests JC5000 said: It's like a mental illness True, but not the way you think. Reply
bit_user hwertz said: I'm not suggesting, at all, that this is a good way to get anything done. But, people pay to put jobs into Pearl (and then the people running the mining rigs get some pay for doing the computations.) And they do have those inference endpoints so you can use this to run real workloads, i.e. it's not some 'theoretically it could be used for work' but no practical example of doing it. But, what is the company's revenue? And are they letting miners run, even when they don't have real jobs that were submitted? To me, that's the real problem. If there's not real work for them to do, then let the miners do something else with their GPUs. However, I fear the company doesn't want to do that, because they want a large pool of compute to attract big clients, and if they let resources drift away, then you have a chicken-and-egg problem. I guess the other reason to let miners keep running is that they probably also want the currency to generate a big market cap. Again, in order to get & stay relevant. Probably, this will turn into just another Ponzi scheme, like we've seen so many times before. After the hype dies down, the company will probably fold, but not before its founders and principals have cashed in their coins and gotten a nice payday. Reply
FunSurfer That "budget GPU" 3060 Ti cost me $1,000 in 2022 Reply
Key considerations
- Investor positioning can change fast
- Volatility remains possible near catalysts
- Macro rates and liquidity can dominate flows
Reference reading
- https://www.tomshardware.com/tech-industry/artificial-intelligence/SPONSORED_LINK_URL
- https://www.tomshardware.com/tech-industry/artificial-intelligence/ai-cryptomining-networks-320-000-rtx-3090-class-gpus-allegedly-burn-112-megawatts-of-power-on-zero-useful-ai-computation-pearls-gpus-are-doing-random-matrix-math-study-claims#main
- https://www.tomshardware.com/subscription
- Amazon says its data centers consume only 0.075% of the water Americans use for watering their lawns and gardens — company also boasts of its improvements in wa
- Snapmaker launches $150,000 Innovation Fund for open source 3D printing — cash rewards target developers backing the U1 toolchanger across Klipper, OrcaSlicer,
- Marvell details vision of optically-interconnected data centers spanning across thousands of kilometers — new interconnects sampling later this year would allow
- U.S. gov't orders Anthropic to disable its newest AI models worldwide due to security threats — ban on Claude Fable 5 and Mythos 5 bars access by any foreign na
- Score 32GB of DDR5 RAM from only $240 in these Newegg hardware bundles for Intel and AMD gaming PC builds — huge savings on premium Gigabyte motherboards couple
Informational only. No financial advice. Do your own research.