Rigario (@Rigario) header
Rigario (@Rigario) profile picture

Founder @holoweave. I mostly post about agents and a bit of investing. Built product orgs at CoinMarketCap, Binance, Chainlink and Zendesk.

Joined April 2012
Photo by Rigario (@Rigario)Photo by Rigario (@Rigario)Photo by Rigario (@Rigario)Photo by Rigario (@Rigario)Photo by Rigario (@Rigario)Photo by Rigario (@Rigario)Photo by Rigario (@Rigario)Photo by Rigario (@Rigario)Photo by Rigario (@Rigario)Photo by Rigario (@Rigario)
After testing Grok Bot, Muse, Hermes Agent, Hark, Instinct, and OpenAI Dots on real personal assistant work, shopping, and agent coordination through Holoweave Weave... Hermes on its own computer/VM is S-tier once you customize it. For the best balance of capability, always-on access, and actually finishing jobs, Grok Bot is my top pick right now.
1
2
1
3
384
Do we think as a society we are well past the point we can do without AI? Or has it become part of daily life?
7
Deals are coming early and Google is taking part with their subscription. Does that usually happen?
We’re kicking off our Black Friday deals today with early access to our best deals for current AI and storage subscribers: Get up to 50 percent off Google AI Plus and annual Pro plans starting today through 11/26, so you can try out our latest models, including Gemini 4 Hydrogen and 4 Neon, AND get more storage! Note that your exact offer will depend on your current plan. Terms apply. Check it out here: one.google.com/ai?utm_source…
8
The axis I feel day to day isn't capability. If you use the same Hermes Profile, ChatGPT stops at the first chance it gets. Qwen 4 Max starts working when I only asked a question. If we keep the harness fixed and swap the model. The temperament shows up fast. Where would your daily driver land on this?
Sharing this diagram again because I need people to understand the difference here Capability is the only thing you see on benchmarks, and it’s not even a complete picture of it
41
Replying to @Teknium
This will be big given Gemini and their current set of models. Came out of nowhere but I'm not complaining.
1
393
I am often asked why @holoweave raised to buy hardware. Its very simple, I either buy hardware or hire people. The hardware ensure that I can get work done for the price of electricity. If I'm already happy with the level of intelligence its not going to get dumber!
Replying to @Rigario
Exactly, too many options for these GPUs! Some people warn me that our current GPU's will be the "Nvidia Tesla V100 12GB" of our times but they don't get that 256-512GB of memory has surpassed a bar of intelligence that even V100's couldn't do easily.
1
43
Indeed, give me the Gemini models in Hermes and I'll upgrade to Ultra immediately.
1
8
1,506
I really hope its not something that will mess up usage even more.
1
2
1,250
Replying to @HandsomeHank_
I'm still on the legacy SuperGrok Heavy plan, and it's just as bad. It's definitely not the new plans I think.
1
1
9
Replying to @HandsomeHank_
I'm not sure it has ever been fixed. And this is after I've used it quire sparingly.
1
1
66
Replying to @krunkosaurus
Exactly, this is not some kind of crypto mining situation where if prices go down, we will stop using the GPUs. The GPUs are going to be used no matter what; there's just too many things to be done.
1
17
953
Replying to @MiaAI_lab
Strange world we live in. Would be much better if their product strategy was a bit more coherent.
95
Replying to @krunkosaurus
Concurrency seems a bit low on the studio right? So even though they match single stream, the sparks should probably do better under parallel load.
1
104
Replying to @JiaShen_AI
Yeah, this is quite similar to what I experienced, where superseded facts do come back once in a while.
14
Replying to @scaling01
Gemini models are legit good this time around. The problem is that we still need to use their harness.
157
Playing with Qwen 4 Max its almost the exact opposite of the ChatGPT models. ChatGPT seems very lazy at the moment, stops at the first chance it gets. Qwen is so eager to do things I can ask a question and it starts working on it. Both on the same Hermes Profile
48
So far the only thing I like about 6.2 Sol and Astra is that we have ultrafast on the $200 plan. Its not smart but it feels good. I've been saying this for abit, speed can make up quite a bit for intelligence.
1
37
Replying to @armalo_ai
Yup agreed, very common failure mode.
1
58
Replying to @shrey_sancheti
Nice, that could slot in as the store instead of git. I also run a pass every night that rereads memory and replaces any fact a newer one contradicts, even when it sits under a different key. Same-key overwrites don't catch that.
2
16
Replying to @HermesWatcher
This is probably the right use for my VM that is included in the subscription. I already have Hermes running on a computer. I don't really need another VM running another Hermes instance.
1
202
Retrieval was never where my agent memory went wrong. It more when old facts sitting next to newer ones that contradict them, and the agent trusting both. What breaks first in yours, finding the fact or keeping it true?
// RAG in a Trenchcoat // This NeurIPS 2026 paper shows that a minimal RAG memory matches state-of-the-art agent memory systems on standard benchmarks. (bookmark it) BRIDGE stores chained observation notes, retrieves them with BM25 plus embeddings, and shows them in time order. It uses one LLM call per write or query and about 5-7K tokens per question. It matches the prior best within confidence intervals on LongMemEval-S and BEAM, and beats the best memory baseline on EverMemBench by 5.7-6.3 points. The authors then test what memory lets an agent do. In their probes, the agent must apply a remembered rule or pattern that the question never mentions. No single design wins all probes. LangMem, which extracts atomic facts, wins rule application. BRIDGE wins pattern induction. Fact extraction loses the shape of a demonstrated pattern, while block summaries retrieve the rule, but the agent often does not apply it. Showing the rule more prominently closes most of that gap. Don't overcomplicate your memory system. Old retrieval methods are still very useful for agents. arxiv.org/abs/2611.22179 Chat with Paper: academy.dair.ai/papers/rag-i…
7
3
100