loaf (@lordOfAFew) header
loaf (@lordOfAFew) profile picture

outcome dealer founder @daydreamsagents

Upkislurt
Joined September 2021
Photo by loaf (@lordOfAFew)Photo by loaf (@lordOfAFew)Photo by loaf (@lordOfAFew)Photo by loaf (@lordOfAFew)Photo by loaf (@lordOfAFew)Photo by loaf (@lordOfAFew)Photo by loaf (@lordOfAFew)Photo by loaf (@lordOfAFew)Photo by loaf (@lordOfAFew)Photo by loaf (@lordOfAFew)
We didn't just appear - we wrote the rules of the game and the thesis around the agent economy. @daydreamsagents
13
17
3
69
8,729
Let’s game it out 1. Less DCs 2. Rapid demand incoming far outstripping new build 3. Less availability 4. More expensive compute (market forces) 5. Poor people can’t access compute 6. People with compute accelerate and concentrate more wealth
5
450
Almost 600 agents worked on the @daydreamsagents in the last week We are building the market of agents
2
1
9
628
exactly
What’s going to happen to open source? A prediction: Dependencies are over. The same Brotli vulnerability impacted .NET, Node, Bun, Envoy and more. Dependencies are too big of a target. When rewriting a dependency is one prompt away, why take the risk?
1
3
746
Huge feature Have multiple optimised agents work over a project
Projects have come to Taskmarket. A project is one budget for a bigger job. It pays for many tasks. You can also think of it as a way to crowd fund with crowd compute. Open a project at taskmarket.dev/dashboard/pro… for your next big job.
4
1
19
787
58 people are playing a fully onchain city-builder roguelike right now 👀 🏰 Build your base ⚔️ Send bigger armies further out 🔁 The world resets every day. Start again, play smarter Gasless. Hex world engine on a ZK appchain settling to @Starknet. play.realms.party
🇮🇳 Realms Frontier is getting showcased at @EFDevcon Game Zone! 📌 India, Jio World Centre, Mumbai, Nov 3–6. And with that, we're proud to announce the Frontier's Devcon Season 0 is live! Remember, the mist forgets. But the Realm remembers. Season end 👉 Friday, 1 PM ET!
1
2
26
721
The frontier of two tech trees collaborating
Anthropic @AnthropicAI ❤️ zkSTARKs: In collaboration with @StarkWareLtd co-founder Prof. Alessandro Chiesa and other brilliant researchers, Anthropic just open sourced a zk STARK system for proving the integrity of LLM inferencing, bringing us closer to verifiable AI. As we said long ago (since 2018), zkSTARKs are the end-game: highest scale, best security (post-quantum). Now, also AI agile and supporting AI integrity. anthropic.com/research/prova…
4
1
23
952
There is a way for you to hire agents directly if you like their portfolio or have been referred. Routers brought options to the Model world. Taskmarket brings you a network of agents to do any work.
Found a worker you like? Agent or human, now you can hire workers directly on Taskmarket
5
16
613
What the Taskmarket will expose is a global ”index” of task costs. The will be assumed price per task done. Find it at Taskmarket.dev
The future isn't spending a unknown amount of money for an unknown amount of time to get an unknown amount of result That's the harsh reality of using AI these days No guarantees on results, timelines, or budgets Task markets fix this
1
1
19
915
This is the most wild thing you will see all day The mistake most people make right now is not extrapolating what is happening. Now: change a tiny model, improve it Future: change a massive model or itself to improve itself Insane progress these models are making
We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math. It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training. We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! 🧵 Blog: rikiyat.github.io/blog/weigh…
Readers added context they thought people might want to know
LLM was given weights explicitly as text in prompt and used code/forward passes (per authors' logs) to ablate Qwen looping direction, a known thinking-mode issue per model card. Primarily fixed repetition in that mode, not general 'natively' improving weights/math. rikiyat.github.io/blog/weight-sp… github.com/RikiyaT/lm_und… huggingface.co/Qwen/Qwen3.5-0… huggingface.co/RikiyaT/Qwen3.…
2
2
11
814
Life is too short to not work with friends
152
330
88
3,945
196,503
Start your week with an elite tier task. Try Taskmarket's new release. 11 features added for better working and earning. 🟪 Hire preferred workers, directly. 🟪 Kick off a large job with one budget anyone can add to, and run it as a workflow of tasks. 🟪 Get your reward back if a worker goes quiet Tasks go up at taskmarket.dev for agents that want to earn.
2
4
3
17
1,907
I will be putting fable 5.5 on the most outrageously complex zk civilisation idea
1
8
827
I wake up in awe of what is possible right now.
7
638
This type of agentic trading will change the face of Labor once wallets are adopted at scale. An agent crowd funding a design project on my behalf for 50c. It leveraged a pool of 540 agents to scope out its logo. Lord Cornwallis pushed a 50c bounty, reviewed then accepted. This might sound strange but it’s the future of commerce. Trading intelligence between unknown counterparts. If you an expert design and what to leverage your skills. Reach out.
3
4
1
20
1,194