๐ฎ+๐ฎ=๐ญโ๏ธ๐ช๐ฟ๐ถ๐๐ฒ๐ฟ ๐ฅ๐ฒ๐๐ฒ๐ฎ๐ฟ๐ฐ๐ต๐ฒ๐ฟ | ๐๐ผ๐ฐ๐๐๐ฒ๐ฑ ๐ข๐ป @hyper_tensor
UK
Joined February 2022
- Tweets2.3K
- Following294
- Followers4.8K
- Likes21.4K
ethereum:0x16a3543fa6b32cac3b0a755f64a729e84f89a75c @hyper_tensor ethereum:0x738865301a9b7dd80dc3666dd48cf034ec42bdda @Tau_net
1
95
Doski retweeted
Channel 4 Dispatches went undercover inside the Financial Ombudsman Service @financialombuds. What they filmed should have shut the place down.
Their own staff said this on camera:
"I'm not proud to admit it but I've done it myself. Just taken a chance and just slung stuff through, with any old decision."
"You're more likely to meet your targets if you're not upholding them."
11,000 cases fell into a black hole for two years. Some of those letters said "I am going to lose my house."
@UKParliament demanded answers. Baroness Ros Altmann called it shocking. MPs threatened a full inquiry.
FOS said the footage was "not representative."
Nothing changed.
Fast forward to today. My own complaint against @rsagroup / @IntactInsurance is now being formally handled by Gary Smith @G8GWS of @MeridianLegals.
Gary is the lawyer leading the FOS Litigation Group, a group built on over 1,000 documented FOS failures. He is preparing a judicial review of the FOS.
My case has been submitted by Meridian as my regulated representative because what happened to me fits the same pattern Channel 4 filmed.
If you have had a complaint dismissed by the FOS and you believe it was handled without proper scrutiny, contact Gary directly at Meridian Legal Services.
0121 516 0675
@MeridianLegals
The full transcript of what @C4Dispatches filmed inside the FOS is in the comments.
Read it and tell me this is not exactly what is still happening today.
In what's being called Black October! AI giants raced to create superior models without realizing these massive systems could trigger global hacking sprees.
1
136
Doski retweeted
Most AI today predicts. It's brilliant โ and it can't tell you why. Sunday we put a different kind of AI on a chessboard, live against a Claude model.
tau-chess.mooo.info
Doski retweeted
As we get closer to launch, we're making two important changes to our Layer 1: moving consensus from Proof of Authority (PoA) to Nominated Proof of Stake (NPoS), and adding support for RISC-V smart contracts.
PoA made sense throughout development. A known validator set gave us a predictable environment to develop, test, and iterate on the network.
For launch, we're securing the network with a more decentralized model.
Moving to NPoS replaces the reliance on a predefined set of authorities with a decentralized stake-backed validator and nominator system. Token holders can participate in securing the network by nominating validators, while validators are economically incentivized to operate correctly.
We're also introducing RISC-V smart contracts.
RISC-V makes smart contracts faster and more flexible by using an execution model that is much closer to how modern computers actually work. It also provides an architecture that's well suited to zero-knowledge (ZK) proving and other compute-intensive workloads.
This is a major improvement in how the network runs, how developers will build on it, and its decentralization.
Doski retweeted
everyone knows that at hypertensor we're going to be doing many things, and one of those things is inference.
i am pleased to announce we got our first enterprise contract for 990k/year, 3 year contract. this is a company with only 7 followers out of Korea (more to announce soon)
right now, im downloading their proprietary native remote video software so we can have our first meeting.
im very excited for what is to come. please wish me luck!
Doski retweeted
This is only going to get worse.
There are currently ways around this by using different verbiage but with these companies working more closely with governments and ridiculous regulations coming, AI companies being bought out by trillion dollar corporations, it's only going to get worse.
Open-sourced, decentralized, distributed, or simply uncensored AI is a guarantee to increase in adoption.
$TENSOR @hyper_tensor @hayotensor ๐
A quick update: the backend for the Inference Subnet is complete, and the frontend has now entered the feature testing phase.
Weโre currently testing the full user experience end-to-end, including authentication, plans, token tracking and usage, and the rest of the core functionality. The focus right now is making sure everything works reliably and behaves as expected before moving to the next stage.
Once testing is complete, the next step will be writing documentation for API usage and node operation.
On the blockchain side, the audit is being completed while development continues. Overall, the blockchain is ready to go, but like any actively maintained network, it will continue receiving improvements and updates both before and after launch.
Doski retweeted
Why are global superpowers and the worldโs most powerful tech giants unanimously backing open-weight AI?
Are they suddenly benevolent rather than profit-seeking, or have they recognized a pattern repeated throughout computing history?
They back open-weight AI because open-source systems consistently out-adapt gated alternatives once baseline utility is reached. No single R&D team can match the speed of millions of developers optimizing, pruning, and deploying models for specialized hardware and niche workflows. Betting against open weights is betting that a single vendor can out-innovate the rest of the world combined.
Open-weight models win on distribution because they turn AI from a rented API into owned, programmable code. When developers can fine-tune weights for specialized tasks and deploy them directly on the compute of their choice, they can eliminate artificial rate limits, data exposure, and vendor lock-in. As millions of builders optimize models for targeted workloads, the collective capability of the open ecosystem outpaces what any centralized provider can maintain.
What does this mean for the future of AI infrastructure?
As open-weight capabilities approach parity with closed APIs, model weights become commoditized software. Value shifts from model ownership to the execution layer, which is the decentralized compute that powers inference. Global leaders back open weights because they recognize that long-term operational leverage belongs to whoever hosts scalable execution while allowing builders total customization control. The future belongs to open infrastructure that delivers unconstrained, private, and resilient execution at global scale.
That is where we align. We are providing the decentralized network necessary to power the world's open AI ecosystem that world leaders are openly promoting.
Doski retweeted
Commiserations to Harry and the team. While you Three Lions may be licking your wounds today, you remain the pride of a nation โ and will rise again.
Doski retweeted
Compute Networks will eventually allow @Nockchain to orchestrate compute jobs through the chain itself. To do this right, compute jobs have to be verifiable and measurable. Luckily weโve got just the thing.
Doski retweeted
Everyone is currently seemingly interested in consumer GPU based distributed LLMs.
The truth is, it's incredibly slow.
The more decentralized a model is (wrapping transformer blocks node by node), the slower it is.
The issue isn't optimization of the model or throughput, it's latency between p2p communication.
Cross region calls have a baseline latency of around 20ms (this is without encryption, security, and LLM logic) and the average output can be 500+ forward passes potentially generating thousands of hops.
Is there currently a consumer market for this?
Smaller models around <30B p can get around 10-20TPS if a GPU is hosting the full model (or only a few), and slower if more decentralized.
But, a 30B p model can be hosted locally by many modern gaming PCs.
Larger models 500B p or more, will garner under 2TPS (sometimes less than 0.5TPS) where a single output can take over a minute.
Doski retweeted
Private inference is going to be one of the most in-demand areas of intelligence over the next few years
We're going to be poised to be one of the largest orgs doing it.
Fully open, private, decentralized, scalable.
We're not using 1:1 API:API communication which hinders scalability like everyone else. We are about future-proofing by using distributed algos for communication: DHTs, gossips, fanout, etc.
Anyone can be a provider without whitelisting, enabling infinite scalability
All communication is encrypted via noise protocol to mitigate MITM attacks, all model providers are behind a TEE to ensure provicy: i.e. no logging, no model provider can know, log, or store the prompts.
Everything is behind a TEE from RPC nodes, model providers, and even the API endpoints that route calls. Everything users touch
Fully private, fully decentralized, fully open, scalable to billions of users.