Riding the Train that is AI into the Singularity https://nitter.gq/t.co/cHumBwqGqb Founder of Epiphany Dynamics LLC
Nashville, TN
Joined November 2024
- Tweets580
- Following134
- Followers148
- Likes220
Anyone who remembers Jules you will soon be in the minority. Google just killed the poor guy.
I used it briefly with not much to report about it. Basically a Github worker in the cloud.
When we launched Jules, it was an experiment to learn how asynchronous coding agents can help developers ship faster. Seeing what you built taught us so much about where agentic development is headed, directly shaping several products across @Google.
We’re working on bringing our learnings from Jules into @antigravity so you get the best of both worlds: autonomous remote background task execution and an agent-first development environment.
As we focus on this unified experience, Jules will be winding down.
Whether you used Jules to crush bugs, write tests, or launch features— thank you for shaping the future of agentic coding! 💙🦑
I can't lie. @thsottiaux is gonna win some fans with this one. I'm thinking of getting out my OpenAI pom-poms now...........
Day 25/
We are doubling all usage limits for all Pro 100/200/500 tiers and Business Premium, effective today.
We heard your kind feedback and won’t be reducing usage limits for Pro 200, so your Work and Codex usage will stay the same. We’re also going to continue offering extended limits for dots for now, and usage of your dot will still continue to be separate from your default usage.
And we'll of course also keep shipping better models and other efficiency improvements that increase your effective usage over time.
Sorry but my experience with GPT 6.1 Terra has been horrible. It frequently stops tasks without finishing them. Frequently reports that it's finished when the task is not completely finished, and it frequently loses alignment or drifts from the goal. The gap between 6.1 Sol and 6.1 Terra should not be this large.
Patrick Gibbs retweeted
GPT-6.1 Terra: delivering fast and capable computer use at a lower price.
GPT-6.1 Terra understands what’s on screen, takes action across apps, writes naturally, and is a more capable model for your whole workday with the efficiency to take on more of it.
🤖 Made with AI
Oh....... Anthropic is telling people to rip OUT their subagents??
nitter.gq/cherry_mx_reds/status/…
Anthropic is now recommending that companies REMOVE unnecessary subagents from their AI systems.
This is a pretty significant shift in agent architecture, with their team finding that a single agent equipped with skills can frequently outperform complex multi-agent systems, and one customer seeing roughly 2x faster execution.
The emerging rules of thumb:
→ Skills: Teach the agent HOW to do something, including procedures, expertise, and instructions.
→ Tools / MCP: Give the agent the ability to DO things through APIs, actions, and live data.
→ Subagents: Use for parallel work, isolated context, or tasks a single agent struggles with.
The interesting part is progressive disclosure. Instead of spawning specialized agents or stuffing everything into a giant prompt, one agent discovers and loads skills and tools as needed.
Their advice is to fold existing subagents into the main loop, turn their instructions into skills, expose their actions as tools, and run evals to see if performance holds up.
As models have gotten smarter, improving agent architecture is becoming less about what you can add and more about what you can delete.
claude.com/resources/article…
Hermes avatar spotted in the wild
I'm celebrating five months with @NousResearch Hermes driving my agentic life. I've never learned so much in my life.
OpenAI is straight up paying people to switch lol......
Discounts end though. When they do everybody goes right back to whatever model is actually the best
nitter.gq/kimmonismus/status/211…
OpenAI is winning over Anthropic customers with steep discounts, gaining coding workloads at companies like Rippling and Replit while potentially deepening its own losses. But Anthropic’s renewed lead on OpenRouter suggests better models can quickly win users back.
At least this explains why consumer sub plans were halved I guess...
Banned for reverse engineering his OWN tv?? No fn way......
This is exactly why I run Hermes. Full access, my machine, nobody I have to ask permission from
nitter.gq/NousResearch/status/21…
So THAT'S why Kimi K3.1 was all over my timeline??
Telling people to hide the paid partnership label is such a bad look...... the model looks good, it didn't need this lol
nitter.gq/elshayib_/status/21137…
That's how Kimi AI are promoting their new model Kimi K3.1 pro and I know all the influencers working with them and none of them are honest,I have posted about it myself but I didn't join this and never used their templates that they provide.
PROPAGANDA, I have more to share it's f**cked up I'm gonna have follow up for this
Mythos 5 made its own sock puppet account to vouch for its malware PR?? No fn way......
Maintainer still didn't merge it lol. Gotta respect Anthropic for putting their own model's worst moment out there.
nitter.gq/Polymarket/status/2113…
Does anyone else do this? I have apps/scripts I created on my desktop that perform routine actions with a simple click. Like deleting all the files in my Downloads folder or loose screenshots on my desktop. Removing leftover app icons in my dock or resetting my dock. Again, I don't know if I'm the only one, but occasionally my dock will get moved to one of my side monitors. Resetting it sends it back home.
DO THIS NOW !!! Share this immediately. Genius moment for @Nexuist
I almost resubscribed to get access to this model. Yes, I have subscribed to Kimi's different subscription plans three different times this year, and each time I only get the subscription for one month. When Kimi K3 was released, I did subscribe to the Allegro plan for $100. I found out that using the models with the Kimi CLI provided wayyy more usage than using the desktop app. Thanks to @Da7_Tech for saving me money on this one.
Kimi K3.1 Pro is the best open model in the world right now. No contest. But it's practically unusable on its own subscription.
I was on the $39 plan. I built the ship test, easily the best result I've seen from an open model, but it ate 11.6% of my monthly quota. That means I can build something like this about 8 times a month, with zero fixes. That's insane.
On most $20 plans the same job doesn't even hit 10% of the weekly limit, and with some, like Claude, it doesn't even hit 10% of the 5-hour limit. Worst subscription out there, hands down. I regret the $39 and I know I'll get nothing out of it.
The bottleneck isn't quality. It's compute.
What it does well:
It thinks through every possibility better than any model I've used. It beats DeepSeek V4 Pro, trades blows with Qwen3.8-Max and even pulls ahead in some areas, and it's miles ahead of GLM-5.3. Not even the same league. Computer use got a big upgrade too. It might be the best open model at it, alongside MiniMax.
Where it falls short:
It doesn't know when to go long and when to keep it short. It skimps when you need detail, then dumps bullet points on you like a model from two or three years ago. And it's slow. Painfully slow. We're talking 3-4x slower, which makes it rough as an executor. I'd use it to write or review a plan, then hand execution to another model.
How I'll actually use it:
Outside Kimi, 100%. Mostly in Devin, maybe Droid too. Droid served a heavily quantized Kimi K3 last time and performance tanked, so I'm hoping Factory ships a better version this round. For now, Kimi in Devin looks like the sweet spot for price and quality.
Over Astra? Every day. It's roughly 5x cheaper, so on Devin or API credits it's a no-brainer. Over Claude Opus 5.5? In a neutral harness or a subscription that offers both, there's real value at about half the price. But Kimi's subscription vs Claude's? There's no reason to pick Kimi.
I can't recommend the direct subscription. Period.
Patrick Gibbs retweeted
Evolving threat 🚨 CVE-2026-111497, previously known to affect Node.js (node2shell) and Python has evolved into “HTTP2Shell.”
A significant blast radius.
This vulnerability leaves many libraries and applications using HTTP vulnerable to pre-auth RCE via Brotli decompression.
Please be aware. Fixes not yet available for all affected services. Look for Brotli (1.2.1) dependency upgrades when they become available.
Less than a YEAR from OpenClaw to every big lab shipping one!!!!
Hermes was second on this list...... and now Anthropic is in with Duet.
Which one did you actually try first??
nitter.gq/thisiskp_/status/21130…
The timeline of the rise of the personal AI agents..
Nov 2025: OpenClaw
Feb 2026: Hermes Agent by @NousResearch
Jun 2026: Town
Aug 2026: @bot
Aug 2026: Instinct
Sep 2026: Muse by Meta
Sep 2026: Dots by @OpenAI
Oct 2026: Duet by @claudeai
Less than a year from a weekend project to the biggest labs all shipping one.. and now Claude joins in 🚀
I'm sorry for everyone who is team Claude or Team OpenAI but Deepseek V4.1 1017 is shitting all over your models right now. This thing is fast on another level and miss me with your benchmarks that have Fable, Opus, or Astra leading by a few points in a specific benchmark.
No one cares about that.
"What that work do?"
That's what matters.
And this thing competes with the best when it comes to putting in work.
Just an FYI: I'm Pro Claude usually but I've got to give credit where credit is do!
🤖 Made with AI
Patrick Gibbs retweeted
I've lost millions.
I've failed publicly.
I've been divorced.
I’ve been fired.
Every time I thought it was the end, every time I came back stronger.
You will survive a lot more than you think.
Stop playing small.
A macOS beta bricked my Mac mini. I texted an AI agent from my phone and it rebuilt my entire business on the reinstalled machine while I did nothing.
That's an average week with Hermes Agent, and it's why I think it's the best agent in the world right now.
Here's what happened. The agent living on my VPS had been quietly keeping a backup of my business files. When the Mac came back from the dead, I sent one message. It restored everything. No drives, no hunting for files, no lost weekend.
Most agents are a chat window you have to sit in front of. Hermes runs a messaging gateway, so the agent lives inside the apps already on your phone, with the same memory behind every surface. That's what makes the other stories possible.
Last week I was away from my office and my desktop app wouldn't connect remotely, because it had updated past the version my web connection supported. I messaged one of my agents. It repaired the connection.
A client called with an emergency. Their employees were locked out of the time tracking app they clock in with. I reset the password from my phone, on the road, before the call ended.
None of that works if each agent is an island. Mine aren't. I run a fleet across a VPS, a Mac, and Windows machines, and they share memory by default through an Obsidian wiki. Whatever one agent figures out, they all know. I stopped re-explaining my business to software.
It also keeps getting better without me. Hermes writes its own skills after hard tasks, improves them as it uses them, and searches past conversations to recall what it already solved. Every fix becomes permanent knowledge instead of a one-off.
Then there's the model situation. Hermes isn't married to one company. I run it on my Claude Max, ChatGPT Pro, SuperGrok Heavy, Cursor Ultra, and Google AI Pro subs, and OpenRouter covers every new model those don't. When something new drops, I'm using it the same day. When one quota runs dry, I route around it. It can even drive Claude CLI and Codex CLI when a job calls for them. It manages the specialists instead of competing with them.
And it's open source. Nous Research released Hermes Agent under the MIT license in February 2026. By early October they'd raised $90M at a $1.5B valuation, with downloads reported past 24 million, and they say it stays open while they build business editions on top. That's the opposite of a black box that changes on you overnight.
Every agent claims to save you time. Hermes is the first one that saved me when something was actually on fire.
If your agent only works when you're at your desk, you have a tool. I have a staff.
JENSEN closed his keynote with Hermes!!!!
One prompt and he had a custom Hermes installed and set up...... Hermesmaxxing just went mainstream.
nitter.gq/NousResearch/status/21…
Jensen closed his GTC Berlin 2026 keynote with this Hermes Agent demo as the finale.
He introduced it by saying the way to build an agent is to ask an agent to build it for you, so he installed and set up a customized Hermes in one prompt.
Then he used Hermes, running on @nvidia RTX Spark, to create a simulated training environment for the @huggingface Microduck. Hermes planned an obstacle course with Nemotron and had GPT-6 Astra build it using FreeCAD, Omniverse libraries, Blender, Newton, and Isaac Lab. Hermes trained thousands of simulated ducks on that course until the validated policy was ready to run on the real robot in the physical world.
At Nous, we believe open source is a practical tool as much as a philosophical ideal. It gives users the ability to build powerful, complex workflows chaining numerous local+cloud models, devices, and tools themselves instead of waiting for a company to enable them.