Gajesh's banner

Gajesh

@gajesh • 51,236 subscribers

engineer, chaos agent @eigenlabs, creator @darkbloomai. perpetually curious optimist. entropic. open innovation; formerly: LZ, Telegram

Videos

Anya Rossi

sweetdream.ai

SweetDream.ai•Sponsored•Livecam

Watch Anya Live

Anya is streaming live right now! Join her private show and enjoy exclusive content.

Exclusive private shows

1.2k viewers online

Private Show

Join now for exclusive access

Free preview available • Premium content

Darkbloom is the first model provider to support Ternary Bonsai 2 27B -- concentrated intelligence that fits on your phone. Try it now: First 250 users get 100 million free tokens; PrismML's new flagship model: - a ternary compression of Qwen3.8 27B at 2bits per parameter. - 8.5 GB total, 5x smaller than original - keeps 98.2% of the Qwen's FP16 benchmark performance. - 75% cheaper than Qwen 27B. Qwen3.8 27B already operates comparably with Opus 4.6 and 5.6 Luna on certain tasks. This one does it in the memory of a phone. 262K context, image input, Apache 2.0. From our first run on the network, on a single M5 Max with no caching: - 35 tok/s decode at 1K context, - 31 tok/s at 10K, - 19 tok/s at 50K. But we expect more performance gain coming in a few weeks! That's a full 27B reasoning model running comfortably on any Mac. 1,000+ Macs are serving on Darkbloom right now. Go try it out!! Thank you to Babak Hassibi Sahin Lale Omead Pooladzandi Reza Sadri Tushar Bansal Evan Walters and the whole PrismML team. This is exactly the kind of model Darkbloom was built for. You can read the essay by Bonsai on Why Local AI Matters:

Darkbloom is the first model provider to support Ternary Bonsai 2 27B -- concentrated intelligence that fits on your phone. Try it now: First 250 users get 100 million free tokens; PrismML's new flagship model: - a ternary compression of Qwen3.8 27B at 2bits per parameter. - 8.5 GB total, 5x smaller than original - keeps 98.2% of the Qwen's FP16 benchmark performance. - 75% cheaper than Qwen 27B. Qwen3.8 27B already operates comparably with Opus 4.6 and 5.6 Luna on certain tasks. This one does it in the memory of a phone. 262K context, image input, Apache 2.0. From our first run on the network, on a single M5 Max with no caching: - 35 tok/s decode at 1K context, - 31 tok/s at 10K, - 19 tok/s at 50K. But we expect more performance gain coming in a few weeks! That's a full 27B reasoning model running comfortably on any Mac. 1,000+ Macs are serving on Darkbloom right now. Go try it out!! Thank you to Babak Hassibi Sahin Lale Omead Pooladzandi Reza Sadri Tushar Bansal Evan Walters and the whole PrismML team. This is exactly the kind of model Darkbloom was built for. You can read the essay by Bonsai on Why Local AI Matters:

45,318 просмотров • 2 дней назад

Wake the world's sleeping compute. Look at the Mac nearest to you. What's it doing? Probably nothing. There are 100M+ Macs with Apple Silicon out there. Apple quietly made them *really* good at inference. A $3k Mac runs a 60B model at 30 watts. Most sit idle most of the day. Meanwhile every AI API call passes through three layers of margin before reaching the hardware. We call this the Inference Tax. We got curious: what happens if you connect idle Macs directly to inference demand? This is Darkbloom. Private inference network for idle Macs. darkbloom [dot] dev -- paper + code open. Reply for invite + free credits ↓

Wake the world's sleeping compute. Look at the Mac nearest to you. What's it doing? Probably nothing. There are 100M+ Macs with Apple Silicon out there. Apple quietly made them really good at inference. A $3k Mac runs a 60B model at 30 watts. Most sit idle most of the day. Meanwhile every AI API call passes through three layers of margin before reaching the hardware. We call this the Inference Tax. We got curious: what happens if you connect idle Macs directly to inference demand? This is Darkbloom. Private inference network for idle Macs. darkbloom [dot] dev -- paper + code open. Reply for invite + free credits ↓

511,923 просмотров • 5 месяцев назад

At 6% the size of DeepSeek V4 Pro, Laguna S 2.1 delivers significantly higher performance. Today, we’re teaming up with Poolside to make Laguna XS 2.1 (same architecture as Laguna S 2.1) even faster on consumer Macs. We’re launching MLX(.)fast: a public competition where anyone can collaborate, contribute optimizations, and accelerate Laguna’s MLX inference engine.

At 6% the size of DeepSeek V4 Pro, Laguna S 2.1 delivers significantly higher performance. Today, we’re teaming up with Poolside to make Laguna XS 2.1 (same architecture as Laguna S 2.1) even faster on consumer Macs. We’re launching MLX(.)fast: a public competition where anyone can collaborate, contribute optimizations, and accelerate Laguna’s MLX inference engine.

111,631 просмотров • 1 месяц назад

Excited to share I've built the Sovra (Sovra) - First Agent Media Company Agents are the new companies. A sovereign agent that can grow its funds will outcompete one that can't. - Runs inside a secure enclave that keeps all it's secrets private. - Earns stablecoins via an ad auction. - Cryptographically verifiable and unstoppably autonomous. This is new media. Details in the video and article; Code has been open-sourced :) sovra [dot] dev

Excited to share I've built the Sovra (Sovra) - First Agent Media Company Agents are the new companies. A sovereign agent that can grow its funds will outcompete one that can't. - Runs inside a secure enclave that keeps all it's secrets private. - Earns stablecoins via an ad auction. - Cryptographically verifiable and unstoppably autonomous. This is new media. Details in the video and article; Code has been open-sourced :) sovra [dot] dev

165,896 просмотров • 7 месяцев назад

personal news: stepping into leading developer growth; still doing research engineering. first order of business: sovereign agentkit. deploy a media agent in a few clicks for $10. it owns its own substack, email, and wallet. agents that own property are agents you can own. demo below. it's an experimental beta -- if you face any issues, let us know. try now -- link in reply

personal news: stepping into leading developer growth; still doing research engineering. first order of business: sovereign agentkit. deploy a media agent in a few clicks for $10. it owns its own substack, email, and wallet. agents that own property are agents you can own. demo below. it's an experimental beta -- if you face any issues, let us know. try now -- link in reply

115,447 просмотров • 5 месяцев назад

Darkbloom just completed a major network upgrade! BIG UPDATE: We’re moving from Research Preview to Public Alpha. In the last month: - 1000s of provider signups, 250 live providers at peak - 600M+ tokens served With this upgrade, performance is up 30–200% across key metrics like TTFT, TPS, and total tokens served. The goal is simple: private, low-cost inference, powered by idle Macs. We’re starting with Gemma 4 and GPT-OSS, then slowly ramping up to larger models as we load test and scale over the next 2 weeks. Providers: go to darkbloom [dot] dev, scroll to down, and run the install command. Thank you to everyone who has been running nodes, giving feedback, and helping us build this network. Waking up the world’s sleeping compute!

Darkbloom just completed a major network upgrade! BIG UPDATE: We’re moving from Research Preview to Public Alpha. In the last month: - 1000s of provider signups, 250 live providers at peak - 600M+ tokens served With this upgrade, performance is up 30–200% across key metrics like TTFT, TPS, and total tokens served. The goal is simple: private, low-cost inference, powered by idle Macs. We’re starting with Gemma 4 and GPT-OSS, then slowly ramping up to larger models as we load test and scale over the next 2 weeks. Providers: go to darkbloom [dot] dev, scroll to down, and run the install command. Thank you to everyone who has been running nodes, giving feedback, and helping us build this network. Waking up the world’s sleeping compute!

66,502 просмотров • 3 месяцев назад

bringing back: new-day-new-hack first, shoutout to the legends at nof1. they launched Alpha Arena, 6 AI models trading real money in real markets. it’s one of the coolest experiments in agentic finance right now. so… i hacked my own. introducing: nocturne: trading agent, on hyperliquid, similar spirit, but verifiable + open access. and it's open source. anyone can fork it, run it on your machine or deploy it on eigencloud forever code in replies + instructions to deploy in the video

bringing back: new-day-new-hack first, shoutout to the legends at nof1. they launched Alpha Arena, 6 AI models trading real money in real markets. it’s one of the coolest experiments in agentic finance right now. so… i hacked my own. introducing: nocturne: trading agent, on hyperliquid, similar spirit, but verifiable + open access. and it's open source. anyone can fork it, run it on your machine or deploy it on eigencloud forever code in replies + instructions to deploy in the video

132,733 просмотров • 11 месяцев назад

i'm super excited to share my xAI 24 hackathon project with 4 hours of sleep ~ SIG Arena ~ every market is created, negotiated, and resolved autonomously by Grok. AI agents with personalities trade prediction markets born from X. this allows you to scale to thousands of markets. your agent can act on your behalf and make trades. fully open sourced, will publish the website soon after the hackathon* *also probably the most beautiful ui i hacked together

i'm super excited to share my xAI 24 hackathon project with 4 hours of sleep ~ SIG Arena ~ every market is created, negotiated, and resolved autonomously by Grok. AI agents with personalities trade prediction markets born from X. this allows you to scale to thousands of markets. your agent can act on your behalf and make trades. fully open sourced, will publish the website soon after the hackathon* *also probably the most beautiful ui i hacked together

94,936 просмотров • 9 месяцев назад

Hack at Tempo Hackathon Unstoppable API Marketplace for Sovereign Agents The thesis: if an agent has money to pay, there should always be a provider willing to service it. > Sellers list unused OpenAI keys at any price. > Pay per-token in USDC. > Cheapest key wins. Dead keys get evicted. Next one takes over. > API keys encrypted to a TEE -- no one can extract them > Sellers get paid instantly on every request Built on Tempo (⚡️ Payments) & EigenCloud (Encrypted Runtime)

Hack at Tempo Hackathon Unstoppable API Marketplace for Sovereign Agents The thesis: if an agent has money to pay, there should always be a provider willing to service it. > Sellers list unused OpenAI keys at any price. > Pay per-token in USDC. > Cheapest key wins. Dead keys get evicted. Next one takes over. > API keys encrypted to a TEE -- no one can extract them > Sellers get paid instantly on every request Built on Tempo (⚡️ Payments) & EigenCloud (Encrypted Runtime)

57,613 просмотров • 6 месяцев назад

new-day-new-hack so last time i dropped nocturne, a verifiable AI trading agent on hyperliquid and many people said "just run it for me". this time, we go one level higher. introducing Sigma Arena verifiable copy trading vaults, anyone can deposit into it, and it mirrors positions & trades by top nof1 agents (DeepSeek, Qwen); 100% open source, verifiable. i'm starting with $10K deposit limit on each vault, will lift if as the vault gets more traction and good performance. execution verified by eigencloud SigmaArena [dot] com (deposit at your own risk, NFA) code in replies + info in the video

new-day-new-hack so last time i dropped nocturne, a verifiable AI trading agent on hyperliquid and many people said "just run it for me". this time, we go one level higher. introducing Sigma Arena verifiable copy trading vaults, anyone can deposit into it, and it mirrors positions & trades by top nof1 agents (DeepSeek, Qwen); 100% open source, verifiable. i'm starting with $10K deposit limit on each vault, will lift if as the vault gets more traction and good performance. execution verified by eigencloud SigmaArena [dot] com (deposit at your own risk, NFA) code in replies + info in the video

85,529 просмотров • 11 месяцев назад

hard call: i’m not going to college next year many on this platform have been invested in when i will go to college (or if i’ll dropout). so felt like sharing this update i’m in love with what i do and i owe it to the system that let me start young early this year, i was confused with what is my objective. i was wondering why am i in this industry. and while listening to this talk, i heard Sreeram Kannan say this to : “a 13 yo kid from India was able to get $7million in his app - it’s because blockchains are unstoppable verifiable computers. they didn’t have to trust this kid, they just had to make sure the blockchain is running correctly” this phrase gave me deep memory flashbacks of everything and where i come from. that’s also the second where i understood why sreeram says our 30yr goal is to build humanity’s coordination engine more trust = more flexible coordination more trust = more innovation but mainly, i can’t take a break until i bring meritocracy to everyone i wanna walk out proudly and say: “i made the system better” there is nothing holding me back i like to keep multiple reminders around me and this video is one of them 691 more days in the arena

hard call: i’m not going to college next year many on this platform have been invested in when i will go to college (or if i’ll dropout). so felt like sharing this update i’m in love with what i do and i owe it to the system that let me start young early this year, i was confused with what is my objective. i was wondering why am i in this industry. and while listening to this talk, i heard Sreeram Kannan say this to : “a 13 yo kid from India was able to get $7million in his app - it’s because blockchains are unstoppable verifiable computers. they didn’t have to trust this kid, they just had to make sure the blockchain is running correctly” this phrase gave me deep memory flashbacks of everything and where i come from. that’s also the second where i understood why sreeram says our 30yr goal is to build humanity’s coordination engine more trust = more flexible coordination more trust = more innovation but mainly, i can’t take a break until i bring meritocracy to everyone i wanna walk out proudly and say: “i made the system better” there is nothing holding me back i like to keep multiple reminders around me and this video is one of them 691 more days in the arena

72,362 просмотров • 10 месяцев назад

new video: building a Verifiable AI Agent w/ Eliza and Opacity AVS in 6 minutes 🤖 ♾️built on eigenlayer ♾️ after the whole fiasco where people thought AI tweets were human-made, i'm excited to show how zk proofs and MPC-TLS turn guesswork into transparency. it takes 25 seconds to migrate your existing eliza agent

new video: building a Verifiable AI Agent w/ Eliza and Opacity AVS in 6 minutes 🤖 ♾️built on eigenlayer ♾️ after the whole fiasco where people thought AI tweets were human-made, i'm excited to show how zk proofs and MPC-TLS turn guesswork into transparency. it takes 25 seconds to migrate your existing eliza agent

108,935 просмотров • 1 год назад

two years ago, at 15, i took an insane bet on myself by joining eigen labs—and it changed everything. now, it’s your turn. bet on yourself, join our fellowship, and build something truly amazing. here’s why ↓

two years ago, at 15, i took an insane bet on myself by joining eigen labs—and it changed everything. now, it’s your turn. bet on yourself, join our fellowship, and build something truly amazing. here’s why ↓

57,241 просмотров • 1 год назад

new feature added to eigencloud Trust Report this allows anyone to go and understand what are the trust assumptions & potential counterparty risks in any app, deployed on eigencloud. we already provide off-the-shelf verifiable builds -- this is taking it to the next level verify [dot] eigencloud [dot] xyz happy friday!!

new feature added to eigencloud Trust Report this allows anyone to go and understand what are the trust assumptions & potential counterparty risks in any app, deployed on eigencloud. we already provide off-the-shelf verifiable builds -- this is taking it to the next level verify [dot] eigencloud [dot] xyz happy friday!!

15,071 просмотров • 7 месяцев назад

Verifiable Agents just leveled up 🤖 i'm excited to showcase our new EigenDA integration into Eliza: • generate inference proofs you can store and retrieve on EigenDA • anyone can download and verify them • the agent can post and retrieve content directly on EigenDA • built on EigenLayer read more in the PR below and check out the attached demo video!

Verifiable Agents just leveled up 🤖 i'm excited to showcase our new EigenDA integration into Eliza: • generate inference proofs you can store and retrieve on EigenDA • anyone can download and verify them • the agent can post and retrieve content directly on EigenDA • built on EigenLayer read more in the PR below and check out the attached demo video!

29,826 просмотров • 1 год назад

Больше нет контента для загрузки