Now that we’ve taken time to refocus and strategize our next move with BrainPlay, we’re expanding our efforts from within ShiftLayer. We’re opening up our AntiEvmMev solution to other Bittensor subnet owners. We have developed a defensive contract that detects MEV sandwich patterns and flips bot profits into losses. We’re actively looking to connect with subnet owners who want to deploy this tech within their subnet. We’ll help you set it up at no cost, and get it implemented for you with the team that out smarted a MEV Bot with a 150 TAO loss. so if you’re dealing with MEV extraction, let’s talk. DM us here and we can discuss further. #TAO $TAO #Bittensor

🟩SUBNET 117 has been sold 🔥 @brainplay_ai explained that they will be returning to the testnet for development and that, in the meantime, #SN117 has been sold to a new team. Shiftlayer (behind Brainplay and the recent redesign of their website) confirmed that it's a team with solid Bittensor experience... 👀 I'm very curious... (also because the price of the alpha is still very low!). Do you have any ideas about the team? 🧐🧐 - - $TAO

Bittensor just had a very busy 7 days One governance primitive. Two ETF rails. Four competition launches. $85K in rewards. A fresh 3.3 billion parameter training run going live on Subnet 3. Here's the full ledger. Templar is dead. Teutonic is the heir. Subnet 3 rebranded after the Covenant founder Sam Dare exited on April 10, dumped ~37K TAO, and publicly accused Const of running "decentralization theater." Teutonic-I ran a 1B parameter model and finished. Teutonic-III is now training live. • 3.29B parameters, 18 layers • Gemma 4 architecture, mixed sliding and full attention • Fresh seed king live on Hugging Face • Anyone can submit a fine-tune that beats it on per-token log-loss • Roadmap: 30B in a month. 1 trillion on the horizon. Const just introduced a way to fire bad subnet owners. BIT-0011 "Conviction." Subnet founders lock alpha tokens. Conviction score decays linearly over 30 days. Highest score owns the subnet. Period. The irony nobody is pointing out. The spec was drafted by Sam Dare himself before he exited. Const's structural fix to the Covenant problem came from the Covenant founder. Oro wasn't alone. Four competition launches in one week. • Oro (SN15). First decentralized arena for AI shopping agents. $85K in rewards. Backed by Crucible Labs, Unsupervised Capital, Savant. • Babelbit (SN59). New incentive mechanism Arena went live. • BrainPlay (SN117). Full relaunch under Shiftlayer AI. AI benchmarking through games and agents. • Metanova (SN68). NOVA nanobodies competition live for decentralized drug discovery. Enterprise is here. • Score (SN44) x Manako x PwC France. Physical AI deployment at enterprise scale. • Targon (SN4) x Intel x PwC Maghreb. First rollout of Bittensor-powered Vision AI confidential compute stack. Venice Uncensored 1.2 also trained on Targon. • BitMind (SN34) x CysecOnline. South African digital forensics firm integrated BitMind deepfake detection. • TrustedStake x Kraken Institutional. Official validator staking partnership. BitGo just unlocked institutional subnet token custody. BitGo and Yuma now let institutions stake AND trade Bittensor subnet tokens inside a qualified custodian. The June 2025 partnership was base TAO only. This one covers the alpha-token layer. Subnet tokens are the leveraged bet on this ecosystem. ~$1B in collective market cap. Individual subnets ran 200 to 400 percent in March when Covenant-72B dropped. Until this week, no regulated path in. Now there is one. Four subnets put real revenue on the table. • Synth (SN50) closed Q1 at $70K MRR. Added HYPE, XRP, WTIOIL to dashboards. • Data (SN13) processed 147K jobs in April through Dataverse CLI. • Resi (SN46) launched its institutional-grade property pricing API. The institutional track kept moving in parallel. • Grayscale and Bitwise both filed TAO ETFs on April 2. SEC decision tracks to August 2026. • TAO Institute launched April 15 with a Subnet Risk Index for institutional investors. • Const spoke at Paris Blockchain Week alongside Bpifrance. Covenant tried to break the network. The community rebuilt it. Community miners restored SN3, SN39, and SN81 entirely from open-source code. Zero founder intervention. 70% of circulating TAO stayed staked through the 38% drawdown. If the network looks like this in a week with a major operator exit, ask yourself what it looks like in a week without one.

.@Shiftlayer_Ai ShiftLayer released the fully redesigned BrainPlay platform, shifting from static academic benchmarks to dynamic, human-centric evaluations through interactive real-world tasks and challenges.

🔐 Security update: Coldkey Swap 🔄 We proactively swapped our coldkey following the btcli / btwallet supply chain issue. Everything is safe on our end, and operations continue as normal. Extrinsic: taostats.io/extrinsic/776742… Thanks to the community for quickly investigating and addressing this. #bittensor $TAO #brainplay #shiftlayer #security

1,840 unique visitors in one week. 33,091,000 total requests in one week. We migrated our domain to Cloudflare last week and just checked the stats. Honestly didn't expect that much traffic so soon. It's a good signal that we're building the right thing at the right time. People are clearly curious about real-world AI benchmarking. We're working hard to bring something big to the community. Built on Bittensor. #bittensor $TAO #brainplay #shiftlayer #benchmark

Manifold Labs is proud to announce our first strategic advisory engagement with @Shiftlayer_Ai, providing expert guidance, mentorship, and operational support to accelerate their growth within the Bittensor ecosystem. ShiftLayer, the owners and operators of Subnet 117 - BrainPlay, are building advanced LLM benchmarking systems through the use of real-world games as a form of competition amongst AI agents. These competitions produce transparent real-world benchmarks, allowing agents to develop human-like abilities such as interpretation, prediction, planning and creativity. These human-like abilities will then be applicable to physical robots as a part of the longer-term vision of BrainPlay. We're excited to help the ShiftLayer team scale sustainably and contribute even more to decentralized AI innovation!

This is exactly why we're building BrainPlay (SN117) on Bittensor. Benchmarks get saturated. Public test sets get memorized. Leaderboards get gamed. BrainPlay benchmarks models through real-world gameplay - starting with Codenames, expanding to more games - where every round is dynamic, adversarial, multi-step, and unpredictable. No static question bank. No fixed test set. No "train on the answers". You can't overfit a live game environment the same way you can overfit a benchmark dataset. And this is where Bittensor becomes powerful. Instead of one lab controlling the benchmark, you have an open subnet where: • models compete continuously • performance is visible on-chain • incentives reward real capability, not benchmark optimization • anyone can plug in a better model and challenge the leaderboard It's decentralized evaluation, not marketing-driven evaluation. SWE-rebench is a great move toward decontaminated testing. But the future is dynamic, adversarial, real-time evaluation - not static test suites. If we want to know which models actually think, strategize, and adapt under pressure, we need environments that evolve. That's what we're building with BrainPlay on Bittensor. #ShiftLayer #Bittensor #benchmark #evaluation #BrainPlay #ArtificialInteligence $TAO

One important thing to highlight: this is decoy-based. We're open to collaborating with any subnet owners or teams interested in exploring this together. #mev #bittensor $tao #shiftlayer #brainplay #subnet

🛡️ MEV-shielding is now live on our dApp We implemented MEV-shielded transactions a while ago, but that alone wasn't enough while EVM-MEV was still effective. After extensive anti-MEV work, we've closed the loop and removed MEV from our subnet. MEV may evolve with new strategies - and we'll keep evolving our defenses alongside it. Try it out and share feedback 👇 shiftlayer.ai/buy-stake #bittensor #shiftlayer #brainplay #alphanomics #mev #dapp

Quick context on why we moved from 4 players → 3 players 👇 With V2.0, clue and guess no longer need to be separate competitions. We're fully model-based now, which shuts down that whole cheating vector entirely. This change will be included in the next update. FYI. #competition #subnet #shiftlayer #brainplay #cheating

🎮 Post-V2.0 question 🎮 V2.0 unlocked support for additional games on the BrainPlay subnet. What types of games or which specific game should we benchmark AI for? 🧠 Strategy 🗣️ Social 🤝 Co-op 🎭 Deception 🎲 Something wild? Ideas welcome 👇 #bittensor #shiftlayer #brainplay #subnet #aibenchmark #GamingCommunity

We're proud to share a simplified roadmap. Everything is progressing in the right direction - and we're excited about what's coming next. 🧵 Text Reasoning (LLM) │ ▼ Vision Text Reasoning (VLM) │ ▼ Vision Text Action (VLA) │ ▼ Real-World Robot Integration #brainplay #shiftlayer #benchmark #modality

With the release of v2.0 and full @TargonCompute integration, BrainPlay is no longer limited to benchmarking just text-based language models (LLMs). We are now evolving into a multi-modal AI benchmarking subnet. #brainplay #shiftlayer #aibenchmark #tvm

🚨 BrainPlay v2.0 is LIVE The most important upgrade yet. Truly fair competition starts now 🧠 github.com/shiftlayer-llc/br… 🧵 TL;DR: 1. @TargonCompute now powers all model inference → serverless, confidential, untouchable 2. No more OpenAI shortcuts → if you’re not uploading a real model, you’re not earning 3. Copycats can’t keep up → your model runs invisible inside TVM 4. Validators benchmark at scale → miners run their own containers; validators can spin up mass evals across all miners simultaneously 5. SDK is already integrated ✅ → if bugs pop up, we report → TVM team patches fast 6. Subnet is now modular → more games, more agents, same core — no breakage 🙏 Massive thanks to @manifoldlabs → custom features, instant fixes, real-time collaboration → this upgrade wouldn’t be possible without you This is the beginning of real model wars. On-chain. Verifiable. Incentivized. #Bittensor #DeAI #Targon #TVM #Shiftlayer #BrainPlay #LLM #AIbenchmarking

📢 Subnet Commitment (TaoFlow Context) TaoFlow changes the game: miners subnet owners now move forward as a single team, aligned around the subnet token. Miners receive 41% of supply — and seeing most emissions not sold since launch is a strong signal 🙌 #bittensor $tao #shiftlayer #brainplay #taoflow

👀 You might've noticed... we quietly shipped something new. We've silently deployed the Miner Leaderboard, and it's now fully live 💻 (More tabs and features coming soon.) Feel free to explore and share your feedback - let's improve it together ⚡️ play.shiftlayer.ai/leaderboa… #brainplay #bittensor $TAO #subnet #sn117 #leaderboard #shiftlayer