Patryk Porebski

281 posts

Patryk Porebski banner
Patryk Porebski

Patryk Porebski

@pporebski

Robotics | AI | Full-Stack. Building: https://t.co/hCaZ1ILryd at @fdotinc

Katılım Kasım 2008
1.5K Takip Edilen170 Takipçiler
Patryk Porebski
Patryk Porebski@pporebski·
Yes. A lot of robotics software pain is not in the model. It is in the middleware, data plumbing, and the long chain of brittle handoffs between collection, processing, labeling, training, and deployment. Coding models can clearly reduce the cost of writing that glue. That helps. But we still need the underlying workflow and interfaces to be explicit, repeatable, and operable when sensors, schemas, and hardware change. That is the part I think robotics is still missing. Not just faster script generation, but shared infrastructure for the boring path from robot data to deployed updates.
English
0
0
2
165
Patryk Porebski
Patryk Porebski@pporebski·
@arian_ghashghai Arguable. Claiming that building software for hardware is a terrible idea right now is like claiming that the phone was only about telephony, not software. The irony here is that it is exactly software that will enable actual robot deployments at scale.
English
0
0
0
72
arian ghashghai
arian ghashghai@arian_ghashghai·
robotics is inherently about hardware, however I'm meeting more and more founders who want to find a software (or just non-hardware) business to build for robotics. thoughts: > software is behind hardware (so this realization is correct, but not unique), and "robot brain" is indeed a hard problem to solve (further out than most think). that being said, I don't think solving robot intelligence as a company that is neither 1) collecting data (either by robot deployment, or other means) nor 2) a true research company like PI makes a lot of sense > Selling dev tools to robotics companies is a horrible business idea right now (sounds smart, but not enough robot deployments + nowhere near the #1 pain point) > the most obvious non-hardware opportunity is in the deployment gap. specifically, imo the demand for businesses in manual labor that want to try robotic solutions *today* I believe is much greater than most people realize, however no robot (humanoid to service bot) is ready to work out of the box (i.e. someone needs to come set them up, teleop, maintain etc). if I were thinking about a business, I would think about doing something that helps old-school, regular-ass businesses put robots into their space tl;dr build stuff that actively puts more robots into the world
English
88
45
562
120.9K
Patryk Porebski
Patryk Porebski@pporebski·
@garrytan @demishassabis Literally yesterday I finished building my own voice layer over GBrain, because I needed it so much. Glad it’s now built in natively!
English
0
0
0
349
Garry Tan
Garry Tan@garrytan·
GBrain just shipped v0.40.0 gives your OpenClaw/Hermes Agent + GBrain a voice agent. It's based on Gemini Live. (Thanks @demishassabis it's amazing) Large context, great tool use, full brain access. Mars is a friend, Venus is your EA. My open source gift to you.
Garry Tan tweet media
English
77
70
967
139.1K
wildiris
wildiris@wildiris19·
Regarding humanoid form-factor robots: • A critical issue, completely lost in the current hype and rush to market, is that the people who will be selling, operating, servicing and maintaining field-deployed robots in the future, will by necessity be the same people that are doing those jobs now as regards to farm, construction, marine, oil-field, logging, or mining equipment. • In other words, any robotic system deployed in the field, that requires the additional technical support of a team of Stanford University engineering graduate students, is a commercial nonstarter. • Or to put it another way, the current crop of humanoid robots are completely neglecting the questions of design for manufacturing, operation, service, and maintenance. All absolutely critical elements for any robotic system to ultimately be commercially viable in the field. • What a field-deployed robot needs to be is modular. Its mechanical construction needs to be based on interchangeable subassemblies. • And its computational architecture should come in the form of pre-programmed bricks or modules connected together using a single shared serial interface to form a system of distributed intelligence. • This form of construction allows for easy manufacture, easy maintenance, and easy service. Programming is not part of this paradigm. If one wants to change some functionality in a robot, just swap in a different module. • The upside of this kind of construction is that this is the level of service, maintenance, and rebuild competency that already exists within the workforce currently employed in the industries of farming, construction, marine, oil-field, logging, and mining.
English
6
13
63
17.7K
Patryk Porebski
Patryk Porebski@pporebski·
That's exactly right. A robot that needs Stanford grads to operate is research. A robot that a field crew can operate, repair, calibrate, and maintain is a product. Few people are thinking about this now. Most are focused on making the thing work, not how to operate it at scale. x.com/pporebski/stat…
Patryk Porebski@pporebski

Introducing rFabric. Most robotics teams spend 80% of their time rebuilding internal tools and brittle data pipelines. We’re changing that. rFabric is the control plane for robot fleets. One platform for connectivity, telemetry, remote operation, and the end-to-end data-to-deployment loop. Focus on robot behavior, not infrastructure.

English
0
0
7
397
Patryk Porebski
Patryk Porebski@pporebski·
@KuphDev @fdotinc Appreciate it! I’ll definitely share more soon - already have something interesting that I can’t wait to show.
English
1
0
1
11
KuphDev
KuphDev@KuphDev·
@pporebski @fdotinc YES! Fleet management for robots. It’s a hard problem given how much variability there is between automated systems, But that just makes it a problem that much much more worthwhile to solve. I would love to follow along and see how you tackle this 🫡
English
1
0
0
31
PRUTHVI GEEDH
PRUTHVI GEEDH@pruthvigeedh_·
@pporebski Great initiative! Building custom infrastructure from scratch is a engineering effort. We are completely aligned with this mission at @Neuracore_AI . Standardizing this layer is exactly what teams need to stop rebuilding the wheel and focus entirely on intelligence.
English
1
0
2
72
Patryk Porebski
Patryk Porebski@pporebski·
Introducing rFabric. Most robotics teams spend 80% of their time rebuilding internal tools and brittle data pipelines. We’re changing that. rFabric is the control plane for robot fleets. One platform for connectivity, telemetry, remote operation, and the end-to-end data-to-deployment loop. Focus on robot behavior, not infrastructure.
English
2
6
39
5K
Humble Lee ⓛ
Humble Lee ⓛ@humblelee71·
@pporebski Exactly, that 80% infrastructure tax is the biggest bottleneck in robotics today. rFabric looks like the control plane the industry has been waiting for. Please see dm
English
1
0
0
58
Patryk Porebski
Patryk Porebski@pporebski·
DevOps -> MLOps -> RobotOps
Čeština
0
0
0
127
Ole Lehmann
Ole Lehmann@itsolelehmann·
A lot of people keep DMing me "how do I actually use Hermes if I’m non-technical" Short answer: you barely touch the technical layer after the initial setup The whole experience comes down to 3 things you actually interact with day to day 1. The messaging app you already use (Telegram, WhatsApp, Discord, etc) You text Hermes the way you'd text a chief of staff. "Remind me to follow up with Mike Thursday." "Draft 3 replies to the emails from yesterday." "Find me the best flight from Berlin to SF on May 16." It just handles it. 2. Slash commands These are one-word shortcuts you type into the chat so you don't have to write a paragraph every time For example, one I use is /goals You type /goals, tell it what you're working on this quarter, and it remembers. Your daily briefing pulls toward those goals. When you ask for advice it factors them in. When it has free cycles overnight it works on them while you sleep. Other useful ones: /schedule for recurring tasks /skills to turn capabilities on and off /memory to peek at what it already knows about you 3. The dashboard A web page that shows everything your agent is doing in one view: > The tasks it's running right now > What it did overnight while you were asleep > The tools it touched (email read, calendar updated, Slack sent) > Token spend per task and per model, so you can see exactly what each workflow is costing you > The skills you have enabled and how often each one fires And a TON more. So when the agent starts to feel like a black box (like OpenClaw did for me tbh), this dashboard gives you the receipts And the thing that actually makes Hermes work for non-technical people is skills Skills are pre-built capabilities you install from a library Inbox triage, calendar management, managing your ads, solving customer tickets, family scheduling, daily briefing, etc You browse, tap install, and the new capability shows up in your chat the next time you text the agent It's the app store moment for personal agents Once you've connected your messaging app, set your goals, and installed 5-10 skills, you have a real personal agent running in your pocket And you never had to open a terminal
Ole Lehmann tweet media
Ole Lehmann@itsolelehmann

HERMES AGENT FOR DUMMIES Everyone on X keeps talking about Hermes Agent and I finally get why: Once you have an AI that's always-on, remembers everything, and you can just text from your phone, you're never going back to a regular AI chat window. That's Hermes. You text it like an executive assistant and it just handles things. Think of it like an affordable OpenClaw that actually works and is reliable. I was using OpenClaw earlier this year, but it kept breaking and the costs were adding up, so I quit personal agents for a while. Then I set up Hermes and it's what I wanted OpenClaw to be from the start. Here's how it works: > You can talk to it on 19+ messaging platforms (I use Telegram and Discord) > You give it a personality file called SOUL .md so it behaves how you want. > You give access to your email, calendar, full browser access, and whatever other tools you use through integrations. > And you plug in whatever models you want as the brain. I use GPT 5.5 for heavy thinking and DeepSeek for lighter tasks so I'm not burning tokens on every little request (under $20/month in token spend). And I run mine on a Hetzner server for about $5/mo, which enables Hermes to be always-on and persistent. So you can schedule tasks that run autonomously, even while you sleep and even while your devices are off. For example, I have one that scans my email for customer support tickets every morning, matches customer to stripe data, sends out missing links, checks refunds etc. Stuff that used to take me HOURS. And the part that actually separates it from OpenClaw and everything else: Hermes writes its own skills from experience. Every time it completes a task, it saves what worked and turns it into a reusable skill it can run again without you explaining anything. So it literally gets smarter and better over time for your specific workflows. Here's what other people are doing with it: > inbox zero and calendar management from Telegram on their phone > overnight coding that debugs itself while they sleep > a family WhatsApp bot where 5 people share one agent to get stuff done > daily briefings texted every day at 8am > smart home control through Home Assistant > etc Once you try texting an AI that already knows your whole setup and just does things when you ask, opening a chat window and starting from scratch every time feels broken. This is the first AI agent that feels like a legit Chief of Staff.

English
33
45
618
66.6K
Kevin Simback 🍷
Kevin Simback 🍷@KSimback·
It doesn’t have to be markdown *OR* HTML, it’s markdown *AND* HTML At least that’s how I do it in my second brain Get the efficiency of .md and the visual benefits of HTML Please don’t go spend the weekend redoing your setups Do this instead - if you already have a big .md repository, just have Claude Code build an HTML interface layer that sits on top It can create nice HTML files that live alongside your .mds and you can work with it all via your personal “second brain” webapp You can keep using Obsidian / md editors the same way, now you just have a second, more visually rich, interface into your second brain
Thariq@trq212

x.com/i/article/2052…

English
19
8
71
10K
Thariq
Thariq@trq212·
HTML is the new markdown. I've stopped writing markdown files for almost everything and switched to using Claude Code to generate HTML for me. This is why.
Thariq@trq212

x.com/i/article/2052…

English
902
1K
12.1K
4.4M
Patryk Porebski
Patryk Porebski@pporebski·
@chesterzelaya great, congrats! I'm curious what's the tech stack behind controling them all from single machine?
English
0
0
1
50
Chester
Chester@chesterzelaya·
three independent drone agents being controlled by one MacBook with no external localization
English
114
331
4.2K
347.6K
Patryk Porebski
Patryk Porebski@pporebski·
@pham_blnh @livekit for sure! I already have full monitoring, telemetry and remote robot operation implemented as part of rFabric. Fully remote, over public network, 30ms latency. Insane. Stay tuned!
English
1
0
0
102
Binh Pham
Binh Pham@pham_blnh·
collect data remotely using @livekit, infer remotely using @livekit excited for what’s to come, handling transport layer for robots should be as easy as setting up a call
English
3
6
29
2.1K
elvis
elvis@omarsar0·
arXiv Papers → LLM Artifacts This is how I keep up with AI research now. It's like having access to the most personalized arXiv feed. Automations run everyday to curate papers based a set of rules and insights. Curated papers are indexed and power the artifacts. Agent convert papers to LLM wikis (based on @karpathy idea), which means insights are indexed and easily searchable and reusable. I feel like LLM Artifacts is the natural evolution to LLM Wikis. It's about making that knowledge actionable. Artifacts are customizable via agents. Artifacts can interact with agents and are dynamic in nature. Anything can be injected into the artifact as needed (insights, components, suggested experiments, action items, etc). I can take action on Artifact items with my agent orchestrator (Electron app). So I can ask questions about any paper and automate experiments in the background right from within the artifact. This is more than a visual. It's not a single prompt. It's several proactive agents coordinating to surface interesting facts, knowledge, and insights that I can act on a researcher. Agents are not just for generating useful artifacts, they are useful to keep learning and staying on the cutting edge of knowledge. Stay tuned for more.
English
35
99
626
59K