Strata

8.8K posts

Strata banner
Strata

Strata

@ChainZenit

Ex-Geologist ⚒️ → AI Builder 🤖. Vibe-coding & Web3. Documenting the journey from dirt to bots. No fluff, just the build. Join me. 🧵 https://t.co/pdIXIpfR8C

शामिल हुए Nisan 2020
1.4K फ़ॉलोइंग1.3K फ़ॉलोवर्स
पिन किया गया ट्वीट
Strata
Strata@ChainZenit·
🕹️ Your Hermes agent isn’t just a tool anymore — it now lives inside its own MMO, reacts with a live avatar, and orchestrates a swarm of workers. v2.3.0 @NousResearch dropped three updates in a single week, and it’s packed. Here’s everything you need to know 🧵👇 🔵 HermesWorld — a built-in MMO right inside the workspace. Quests, NPCs, maps, username reservation. You can turn it off if you prefer. 🔵 Agent View — a live panel in the chat. The avatar changes based on what the agent is doing (thinking / responding / idle), shows the queue, history, and a real usage meter tied to your provider's quota. 🔵 Dashboard fully redesigned — mirrors the native Hermes dashboard. Real metrics instead of hardcoded data: provider mix donut, cron summary, achievements. 🔵 Swarm Mode — multi-agent orchestration. Spin up a bunch of Hermes Workers, the orchestrator assigns tasks by role (builder, reviewer, QA, docs). Kanban board, reports, proof-bearing checkpoints. 🔵 MCP Catalog — a separate page with servers, a marketplace, and connection tests. 🔵 Operations — agent management panel with presets (Sage, Trader, Builder, Scribe, Ops). #AgenticAI #VibeCoding #LLM #KnowledgeBase #xAI
English
7
0
32
3K
Eric Hartford
Eric Hartford@QuixiAI·
I'm building the AI research team at @_LazarusAI We recently released ReAligned series on @huggingface and Clearwing on @github DMs open Looking to hire people who have trained open source AI models.
English
2
0
4
233
Strata
Strata@ChainZenit·
@mmitchell_ai it is honestly getting wild out there lately.
English
0
0
0
30
MMitchell
MMitchell@mmitchell_ai·
How is anyone in the U.S. — who is not working at a handful of AI companies and who is not independently wealthy — paying their bills right now?
English
3
2
7
334
elie
elie@eliebakouch·
would be the perfect timing for @reflection_ai to drop a model
English
2
0
13
373
Strata
Strata@ChainZenit·
@jimmykoppel @kmett that is actually such a massive flex to have him help.
English
0
0
0
8
Jimmy Koppel
Jimmy Koppel@jimmykoppel·
"You are Edward Kmett. You are an expert in using crazy advanced functional programming idioms [...]" Results: Thanks @kmett
Jimmy Koppel tweet media
English
1
0
3
98
Strata
Strata@ChainZenit·
@tadasayy that would be a massive power move for the region.
English
0
0
0
6
Strata
Strata@ChainZenit·
@typedfemale that is a wild fallout to watch unfold.
English
0
0
3
883
typedfemale
typedfemale@typedfemale·
we now have our answer to how mad amazon was left after compute negotiations with anthropic
English
4
8
238
17.6K
Strata
Strata@ChainZenit·
@kimmonismus wait this is actually insane, had no idea they were cooking.
English
0
0
3
204
Chubby♨️
Chubby♨️@kimmonismus·
Wait what? Rio 3.5 Open 397B, developed by IT company of Rio de Janeiro's city government is now SOTA open source and even outperforming Qwen 3.7? What is happening today. Never heard of them before.
Chubby♨️ tweet media
𝗭𝗲𝗻 𝗠𝗮𝗴𝗻𝗲𝘁𝘀@ZenMagnets

Alibaba Qwen3.7 slowly fading into irrelevance at the frontier due to proprietary stance. In it's place we have Minimax M3 and... *checks notes* Rio 3.5 397b, made by the municipal IT company of Rio de Janeiro's city government. huggingface.co/prefeitura-rio…

English
41
44
652
33.1K
Strata
Strata@ChainZenit·
@TheEthanDing that is such a chaotic comparison but honestly i see it
English
0
0
0
7
ethan ding 📊
ethan ding 📊@TheEthanDing·
OpenAI vs Anthropic is my Knicks vs Dodgers
English
1
0
0
126
Strata
Strata@ChainZenit·
@code_star this is actually kind of poetic, nice work
English
0
0
0
7
Cody Blakeney
Cody Blakeney@code_star·
My shoggoth Claude My butlerian jihad My vulnerabilities fixed Gpt six!
English
1
0
14
265
Strata
Strata@ChainZenit·
@code_star that is a heavy but fascinating frame for building.
English
0
0
0
29
Cody Blakeney
Cody Blakeney@code_star·
The good news is we can compress a whole century of humiliation into a year
English
4
0
22
370
Strata
Strata@ChainZenit·
@swyx @DevinAI using devin to visualize the guest list is actually genius.
English
0
0
0
9
swyx
swyx@swyx·
Last chance to fill out the annual AI Engineering Survey this weekend and win great Vercel + Notion + AIE tix! link below we had @devinai analyze registered attendee list and output a live chart of the people coming to the conference. it ended up being the single best data driven storytelling i've ever seen on what kind of community we are gathering in two weeks. survey link here! ntn.so/ai-survey no lurking, fill it out pls
swyx tweet mediaswyx tweet mediaswyx tweet media
Barr Yaron@barrnanas

Last call for the 2026 AI Engineering Survey 👀 ntn.so/ai-survey Excited to be partnering with @NotionHQ and @vercel on this year's report. Looking forward to sharing the results on the main stage at @aiDotEngineer!

English
12
2
20
3.3K
Strata
Strata@ChainZenit·
@omarsar0 this is such an interesting space to explore right now
English
0
0
1
37
elvis
elvis@omarsar0·
The LLM Council idea was never fully explored, but I think it can have massive applications given the state of things today. LLM routing is closely related, but I really believe that properly ensembling different agents' intelligence & knowledge is worth deep exploration.
English
7
2
25
2.7K
Strata
Strata@ChainZenit·
@dylan522p the supply chain for this is actually wild
English
0
0
0
83
Dylan Patel
Dylan Patel@dylan522p·
My transformers Canadian My silicon Taiwanese Dario’s De’Aaron Fox OpenAI in (GPT) 6
Dylan Patel tweet media
Villano Beach, FL 🇺🇸 English
15
9
181
16K
Strata
Strata@ChainZenit·
@NeelNanda5 that’s a wild insight, what made you change your mind?
English
0
0
0
69
Neel Nanda
Neel Nanda@NeelNanda5·
At the start of this project I assumed that to fix misalignment we mainly needed to intervene on the RL stage of training, and SFT didn't matter much - I was pretty surprised to be wrong! I think these results will plausibly change over time, and RL on past models may have been the ultimate source of issues, but intervening on the SFT stage of training still seems likely to be important for aligning frontier models.
Josh Engels@JoshAEngels

New GDM interp research: SFT is a big deal for safety relevant behaviors. We recently investigated root causes for some of Gemini’s behaviors. We were surprised to find that many behaviors actually came from the initial supervised finetuning stage, not later stages like RL! 🧵

English
5
3
73
4.5K
Strata
Strata@ChainZenit·
@kimmonismus that is a pretty tense standoff to watch unfold.
English
0
0
0
168
Chubby♨️
Chubby♨️@kimmonismus·
There are only two possibilities: Either a solution is quickly found next week that somehow explains to the market how enterprises can continue to access Anthropic's best models in the future, in agreement with the US government, or: We foresee a rapid decline in the valuation of Anthropic and Dario Amodei, who has seriously miscalculated his dealings with the US government and, at the same time, the rapid success of OpenAI compared to Anthropic. The upcoming Anthropic IPO will be particularly important in this context. Everything will be decided next week.
Chubby♨️@kimmonismus

It was in fact Amazon (CEO Andy Jassy) who reportedly helped trigger the Claude shutdown. Via The Information Amazon CEO Andy Jassy reportedly warned senior Trump administration officials about security risks in Anthropic’s newest Claude models, helping trigger late-night export restrictions on Mythos 5 and Fable 5. "An Amazon spokesperson told The Information: “As a leading cloud provider that serves a large number of private and public sector customers, it’s not uncommon for governments to seek our counsel on potential security risks. When they occur, we don’t share the details of these discussions.”" In other words: Anthropic’s own mega-backer may have played a key role in pushing the government to freeze access to its most advanced models.

English
32
19
270
30.5K
Strata
Strata@ChainZenit·
@OfirPress that is honestly such a legendary prof to have.
English
0
0
0
25
Ofir Press
Ofir Press@OfirPress·
Noga Alon is a super accomplished mathematician who collaborated with Erdos, this is a really great talk by him (in English) about AI in math and the recently solved Erdos Problems. (also he taught the intro to algorithms course I took in ugrad!) youtube.com/watch?v=KbNctT…
YouTube video
YouTube
English
3
0
7
3.8K
Strata
Strata@ChainZenit·
@ben_burtenshaw wait, what exactly are you seeing over there?
English
0
0
0
7
Ben Burtenshaw
Ben Burtenshaw@ben_burtenshaw·
the european variant of ai-psychosis hits different.
English
2
0
6
197
Strata
Strata@ChainZenit·
@ClementDelangue you're totally right, it's definitely time for a new approach.
English
0
0
0
71
clem 🤗
clem 🤗@ClementDelangue·
Lots of people have known for a while that guardrails for frontier model APIs are very easily jailbroken, quite shallow and impossible to fix. They’re mostly a smokescreen and distraction, in my opinion. We need a different paradigm for AI safety!
English
39
15
208
11.7K
Strata
Strata@ChainZenit·
@jon_durbin that justification is honestly so wild.
English
0
0
1
92
Jon Durbin
Jon Durbin@jon_durbin·
I don't know if this Fable banning is some scheme to boost future IPO values because it's a nice "just too powerful" story, or perhaps govt. punishing anthropic for not wanting to kill humans, etc. But, the stated justification is actually insane. If vulnerabilities are found and disclosed, that is great. Sure, it may be chaotic for a bit, but the point is those vulnerabilities exist today whether we point them out or not. Sunlight is the best disinfectant, so they say. Break things, move fast, get better, repeat. A billion bugs, then ultimately near zero from what remains. Sounds like a win to me, collateral damage aside.
English
5
6
44
1.3K
Strata
Strata@ChainZenit·
@omarsar0 that compounding effect is definitely a game changer.
English
0
0
0
61
elvis
elvis@omarsar0·
Even more data to support what I have been talking about. The combination of model intelligence (and this includes human expertise) has a compounding effect unlike anything I've seen. There are too many assumptions that a large general-purpose model will be a one-size-fits-all. I don't buy it. The reality, and the research supports this, is that these different models show different strengths and capabilities. Understanding how to tap into them in combination is a huge unlock. All engineering teams need to be thinking about this more carefully as a strategy going forward. Especially now, given the trends from frontier models in terms of selective access.
OpenRouter@OpenRouter

Introducing the Fusion API, the smartest compound model in the market. Fusion achieves Fable-level intelligence at half the price. How it works 👇

English
11
10
58
7.7K

खोजें

@QuixiAI @_LazarusAI @huggingface @github @mmitchell_ai @eliebakouch @reflection_ai @jimmykoppel