#  @_joe_harris_ Joe Harris Joe Harris posts on X about if you, business, matter, investment the most. They currently have [---] followers and [--] posts still getting attention that total [------] engagements in the last [--] hours. ### Engagements: [------] [#](/creator/twitter::4426889653/interactions)  - [--] Week [-------] -71% - [--] Month [---------] +305,594% ### Mentions: [--] [#](/creator/twitter::4426889653/posts_active)  - [--] Week [--] +79% - [--] Month [--] +233% ### Followers: [---] [#](/creator/twitter::4426889653/followers)  - [--] Week [---] +117% - [--] Month [---] +163% ### CreatorRank: [-------] [#](/creator/twitter::4426889653/influencer_rank)  ### Social Influence **Social category influence** [stocks](/list/stocks) #2461 [technology brands](/list/technology-brands) 12.05% [finance](/list/finance) 8.43% [social networks](/list/social-networks) 4.82% [automotive brands](/list/automotive-brands) 4.82% [vc firms](/list/vc-firms) 1.2% [celebrities](/list/celebrities) 1.2% **Social topic influence** [if you](/topic/if-you) 6.02%, [business](/topic/business) #701, [matter](/topic/matter) 3.61%, [investment](/topic/investment) 3.61%, [accenture](/topic/accenture) #4, [the world](/topic/the-world) 3.61%, [tesla](/topic/tesla) 3.61%, [ai](/topic/ai) 2.41%, [software](/topic/software) 2.41%, [build](/topic/build) 2.41% **Top accounts mentioned or mentioned by** [@alloyrobotics](/creator/undefined) [@ananthkashyap99](/creator/undefined) [@scobleizer](/creator/undefined) [@baslyasma](/creator/undefined) [@chrisjpaxton](/creator/undefined) [@doctesla83](/creator/undefined) [@sethbrkv](/creator/undefined) [@madietlx](/creator/undefined) [@astridwilde1](/creator/undefined) [@mattsaniii](/creator/undefined) [@laurisalonen](/creator/undefined) [@hyperlogistix](/creator/undefined) [@kayoum](/creator/undefined) [@rileybrown](/creator/undefined) [@razroochief](/creator/undefined) [@beffjezos](/creator/undefined) [@viraj__acharya](/creator/undefined) [@brennanwoodruff](/creator/undefined) [@5dsolarshift](/creator/undefined) [@dirtman](/creator/undefined) **Top assets mentioned** [Accenture (ACN)](/topic/accenture) [Tesla, Inc. (TSLA)](/topic/tesla) [Alphabet Inc Class A (GOOGL)](/topic/$googl) ### Top Social Posts Top posts by engagements in the last [--] hours "This is why foundation models matter for robotics. Enough compression of reality and the model generalises to things it's never seen. The question is just: how much compression is enough" [X Link](https://x.com/_joe_harris_/status/2018910351090434512) 2026-02-04T04:51Z [---] followers, [--] engagements "@BaslyAsma @kayoum_ Very cool" [X Link](https://x.com/_joe_harris_/status/2020673319494709556) 2026-02-09T01:37Z [---] followers, [---] engagements "@chris_j_paxton step 4: speed up the video 2x and add lo-fi beats" [X Link](https://x.com/_joe_harris_/status/2020719792869777835) 2026-02-09T04:41Z [---] followers, [--] engagements "@BaslyAsma @kayoum_ The investment in the making the product just work. We share the same philosophy. Think there could be interesting ways to collaborate in the future" [X Link](https://x.com/_joe_harris_/status/2020976936433090606) 2026-02-09T21:43Z [---] followers, [--] engagements "@rileybrown hey @grok is this true ;)" [X Link](https://x.com/_joe_harris_/status/2021045123531014507) 2026-02-10T02:14Z [---] followers, [---] engagements "@DavidSacks @PalmerLuckey Polymarket about to go wild" [X Link](https://x.com/_joe_harris_/status/2021052240740614449) 2026-02-10T02:42Z [---] followers, [---] engagements "Models are compression of reality. They become a lens through which you can project a representation of reality. It actually "understands" in that it has compressed the reality of what a toy on top of carpet acts like so the n+1 example doesn't need to be told exactly how to do it. https://twitter.com/i/web/status/2018910339426058557 https://twitter.com/i/web/status/2018910339426058557" [X Link](https://x.com/_joe_harris_/status/2018910339426058557) 2026-02-04T04:51Z [---] followers, [--] engagements "spoiler: it works like magic on alloy" [X Link](https://x.com/_joe_harris_/status/2019611965736522165) 2026-02-06T03:19Z [---] followers, [--] engagements "Prediction: Humanoid robots will offshore jobs before they replace them. Autonomy breaks down on edge cases like the weird box the fallen shelf the unexpected obstacle. You can't ship a robot that fails 20% of the time and the cheapest fix isn't better AI it's a $3/hr worker on a VR headset. So "fully autonomous humanoids" will really mean massive teleoperation factories. Thousands of people in the Philippines clocking in putting on headsets remotely doing physical labor for Western warehouses. BPO for bodies instead of call centers. Huge opportunity for builders in this space. What does" [X Link](https://x.com/_joe_harris_/status/2020617043612192817) 2026-02-08T21:53Z [---] followers, [---] engagements "rubber duck debugging has evolved https://t.co/ZvJGTl7Ret https://t.co/ZvJGTl7Ret" [X Link](https://x.com/_joe_harris_/status/2020709088720826682) 2026-02-09T03:59Z [---] followers, [---] engagements "@Scobleizer @razroo_chief Given few details coming out simplest answer is post merger RIF. Not in anyones interest to make a big deal out of it" [X Link](https://x.com/_joe_harris_/status/2021440242574831888) 2026-02-11T04:24Z [---] followers, [----] engagements "@DocTesla83 @Scobleizer @razroo_chief All the random staff Its far broader than [--] cofounders" [X Link](https://x.com/_joe_harris_/status/2021499223490036070) 2026-02-11T08:19Z [---] followers, [--] engagements "@SethBRKV Better analogy is what if he waited till there was a printing press before selling books on the internet Everyone repetitively building the same foundational things is poor use of resources and human ingenuity" [X Link](https://x.com/_joe_harris_/status/2021683505466683485) 2026-02-11T20:31Z [---] followers, [---] engagements "@madietlx Never been more accessible But still so much opportunity ahead. I had the same experience w EE at uni too" [X Link](https://x.com/_joe_harris_/status/2021716295893590392) 2026-02-11T22:41Z [---] followers, [---] engagements "@astridwilde1 you're onto something https://x.com/astridwilde1/status/2019876567875547548 I'm in the market for a cofounder Looking for someone who has experience in robotics esp dexterous manipulation who is excited by flipping a switch to fix the data bottleneck and unlocking the internet for training to free humans from menial labor THIS DECADE DMs are open https://x.com/astridwilde1/status/2019876567875547548 I'm in the market for a cofounder Looking for someone who has experience in robotics esp dexterous manipulation who is excited by flipping a switch to fix the data bottleneck and" [X Link](https://x.com/_joe_harris_/status/2021747258413490193) 2026-02-12T00:44Z [---] followers, [---] engagements "@benvspak Ingest raw robotics data and get root cause analysis 🤖 https://www.usealloy.ai/ https://www.usealloy.ai/" [X Link](https://x.com/_joe_harris_/status/2021827215189848099) 2026-02-12T06:02Z [---] followers, [--] engagements "Handcrafted goods are expensive because they're hard to make. Handcrafted software is about to mean the same thing" [X Link](https://x.com/_joe_harris_/status/2017742949241745626) 2026-01-31T23:33Z [---] followers, [--] engagements "Getting a robot to 90% reliability is a demo. Getting it to 99% is a product. Getting it to 99.99% is a business. Each step is harder than the last. Most companies are stuck between demo and product" [X Link](https://x.com/_joe_harris_/status/2018167862339461201) 2026-02-02T03:41Z [---] followers, [--] engagements "If you want to go beyond 99% use @AlloyRobotics" [X Link](https://x.com/_joe_harris_/status/2018167874343571608) 2026-02-02T03:41Z [---] followers, [--] engagements "People thought LLMs were nice little toys. They never appreciated. We asked what if we 1000000x training and data volume That's incomprehensible because it would cost a billion dollars. Turns out it's very well worth paying a billion dollars. I think the same thing happens again in robotics" [X Link](https://x.com/_joe_harris_/status/2018542446196691436) 2026-02-03T04:29Z [---] followers, [--] engagements "letting opus [---] cook on TBs of robot data" [X Link](https://x.com/_joe_harris_/status/2019611674479906901) 2026-02-06T03:18Z [---] followers, [---] engagements "Robotics is the final frontier of knowledge work. We worked with @Ginkgo to connect GPT-5 to an autonomous lab so it could propose experiments run them at scale learn from the results and decide what to try next. That closed loop brought protein production cost down by 40%. https://t.co/udKBKxnKlW We worked with @Ginkgo to connect GPT-5 to an autonomous lab so it could propose experiments run them at scale learn from the results and decide what to try next. That closed loop brought protein production cost down by 40%. https://t.co/udKBKxnKlW" [X Link](https://x.com/_joe_harris_/status/2019620042892275781) 2026-02-06T03:51Z [---] followers, [---] engagements "robotics is the the irony of automation its meant to replace labour but getting there will create a decade of new jobs like teleoperation data labelling deployment support before full autonomy arrives we want to help speed up that transition" [X Link](https://x.com/_joe_harris_/status/2020087969131163790) 2026-02-07T10:51Z [---] followers, [---] engagements "@rileybrown telegram did a great job. bot architecture + threads makes it easy to have multiple agents" [X Link](https://x.com/_joe_harris_/status/2022241618011844941) 2026-02-13T09:29Z [---] followers, [---] engagements "Accenture and similar companies have banked billions rolling out other peoples softwares (eg SAP). Big companies will pay for high quality results physical services are a billion dollar industry and now they can be delivered at scale w autonomy. Many of those are training and quality control operations already with high variability output humans (w high turn over). https://twitter.com/i/web/status/2022274501787152545 https://twitter.com/i/web/status/2022274501787152545" [X Link](https://x.com/_joe_harris_/status/2022274501787152545) 2026-02-13T11:39Z [---] followers, [--] engagements "@ananthkashyap99 May be able to help - DMed" [X Link](https://x.com/_joe_harris_/status/2022297112701223097) 2026-02-13T13:09Z [---] followers, [---] engagements "I left a $600M ARR company to build robots. But after speaking with 30+ robotics founders they all confessed the #1 reason robots fail: Data overwhelm. Its costing the industry billions and here's why we raised $4.5m to change that. 🧵" [X Link](https://x.com/_joe_harris_/status/2001920770289844348) 2025-12-19T07:41Z [---] followers, [----] engagements "If I had more time I would have written a shorter letter. [----] version: If I had more time I would have built simpler software. Product management is the art of the writing the least amount of code for the greatest benefit to your users. Product management is the art of the writing the least amount of code for the greatest benefit to your users" [X Link](https://x.com/_joe_harris_/status/2018172216043778531) 2026-02-02T03:58Z [---] followers, [---] engagements "The pattern repeats. 'It'll never work at scale' is always wrong when someone is willing to spend enough. Robotics data is where LLM training was [--] years ago. Wildly underinvested relative to what's possible" [X Link](https://x.com/_joe_harris_/status/2018542457945010628) 2026-02-03T04:29Z [---] followers, [--] engagements "Ironically making robots that replace labour will first require a massive wave of human labour. Robotics learning is emergent. No single person teaches a robot how the world works. Capability comes from many interactions layered together. Community is how that scale happens. Robotics learning is emergent. No single person teaches a robot how the world works. Capability comes from many interactions layered together. Community is how that scale happens" [X Link](https://x.com/_joe_harris_/status/2019287693021389100) 2026-02-05T05:51Z [---] followers, [---] engagements "Building the thing that replaces human work requires more human work than anyone expects. Teleoperation data labelling deployment support there's a decade of jobs here before full autonomy" [X Link](https://x.com/_joe_harris_/status/2019287930347614585) 2026-02-05T05:52Z [---] followers, [--] engagements "99% of the robotics companies that will exist in the next [--] years aren't here yet Its time to accelerate 🥂" [X Link](https://x.com/_joe_harris_/status/2021064637996019730) 2026-02-10T03:32Z [---] followers, [---] engagements "2026 is a terrible time to build a robotics company. Imagine building a software company without AWS Stripe or GitHub. Everything from scratch. That's robotics today. Most robotics teams waste their first [---] months rebuilding the same data infrastructure telemetry internal tools pipelines. None of it is core IP. I started @AlloyRobotics to SPEEDRUN all of it. A modern stack so robotics teams can iterate 10x faster than their competition. If you're building in this space lets talk. https://twitter.com/i/web/status/2021431598823113177 https://twitter.com/i/web/status/2021431598823113177" [X Link](https://x.com/_joe_harris_/status/2021431598823113177) 2026-02-11T03:50Z [---] followers, 19.1K engagements "The Accenture of robotics doesn't exist yet. Someone will build a billion-dollar services business just deploying other people's robots" [X Link](https://x.com/_joe_harris_/status/2022239313833140340) 2026-02-13T09:19Z [---] followers, 81.4K engagements "@Viraj__Acharya Big opportunity" [X Link](https://x.com/_joe_harris_/status/2022272446955721185) 2026-02-13T11:31Z [---] followers, [---] engagements "@ananthkashyap99 Keen to know more" [X Link](https://x.com/_joe_harris_/status/2022292291898282039) 2026-02-13T12:50Z [---] followers, [----] engagements "@MattSaniii Some will. Same answer to why doesnt every OEM vertically integrate today. Cash flow and specialisation. Sell today for positive margin or run on a financing and payback model over time. Well see both again" [X Link](https://x.com/_joe_harris_/status/2022470782702714882) 2026-02-14T00:39Z [---] followers, [---] engagements "The #1 thing slowing down the age of robotics: There is no "RobotOps" stack built for multimodal high-frequency data" [X Link](https://x.com/_joe_harris_/status/2000842090276941873) 2025-12-16T08:15Z [---] followers, [--] engagements "the companies winning in robotics have faster learning loops than competitors. find failure mode understand root cause fix model deploy repeat. if you can compress that loop from weeks to days and you 10x your velocity" [X Link](https://x.com/_joe_harris_/status/2000901230638588128) 2025-12-16T12:10Z [---] followers, [--] engagements "Robots produce 1GB of data per minute. That's 1000000x more than a web app. None of the web-based infrastructure tools (Datadog Amplitude Honeycomb) were built to work with these data formats simultaneously. This is the actual bottleneck to robotics scaling" [X Link](https://x.com/_joe_harris_/status/2001166595432046663) 2025-12-17T05:44Z [---] followers, [--] engagements "@roboticss_alpha @Figure_robot BMW partnership was the real step-function change for Figure if they can scale that deal to the rest of the factory + move into other automotive companies they might have a much stronger GTM motion than any other humanoid company right now" [X Link](https://x.com/_joe_harris_/status/2001594482916233364) 2025-12-18T10:04Z [---] followers, [--] engagements "Most missions never get analysed at all. Not because teams don't care. Because manual analysis takes too long: "Manually replay [--] hours of footage" "Check the external hard drive from March" "Run Mark's Python script. wait which version" So you're forced to choose: Which failure is worth investigating" [X Link](https://x.com/_joe_harris_/status/2001920969884209331) 2025-12-19T07:42Z [---] followers, [--] engagements "Here's why robotics companies stay small: Manual analysis doesn't scale. Either you analyse every mission (need [--] engineer per robot) Or you skip most missions (miss 99% of learning opportunities). Either way you can't scale humans AND improve robots. The unit economics just don't work" [X Link](https://x.com/_joe_harris_/status/2001921034354856101) 2025-12-19T07:42Z [---] followers, [--] engagements "But to go from 99% to 99.999% reliability Requires learning from every edge case. But if you only analyse the obvious failures. You miss the edge cases buried in "successful" missions. Your learning velocity stays stuck" [X Link](https://x.com/_joe_harris_/status/2001921100721349046) 2025-12-19T07:42Z [---] followers, [--] engagements "Why is this fixable now LLMs can reason about multimodal data. You can compress an entire mission with various data types likes camera streams telemetry logs into a report that surfaces: What happened Which moments matter What to investigate first Automatically. Within hours not days" [X Link](https://x.com/_joe_harris_/status/2001921166936805806) 2025-12-19T07:42Z [---] followers, [--] engagements "That's what we built at @alloyrobotics. Now you can analyse every mission. Upload your data. Get a report that surfaces the moments that matter. No manual replay. No Python scripts. Fast learning = fast iteration = better robots" [X Link](https://x.com/_joe_harris_/status/2001921178928394670) 2025-12-19T07:42Z [---] followers, [--] engagements "@IamKyros69 [----] will be the year of more pilot programs. Mass deployment is still a few years out minimum" [X Link](https://x.com/_joe_harris_/status/2003674701789536282) 2025-12-24T03:50Z [---] followers, [--] engagements "Robotics companies spend millions collecting data but its hard to answer basic questions about your robot's performance. After running [---] robot missions you should be able to ask: "Has this failure happened before" "What patterns am I missing" "Where are the gaps in my testing" But answering any of those means manually reviewing 100s of missions" [X Link](https://x.com/_joe_harris_/status/2004692474946851219) 2025-12-26T23:15Z [---] followers, [--] engagements "This is "the complexity threshold" - the moment you have too much data to understand without tools. Most teams hit it around 50-100 missions. After that you're either: Guessing Hiring more engineers Missing patterns that cost you months Or you automate the querying & bucketing layer entirely with @AlloyRobotics" [X Link](https://x.com/_joe_harris_/status/2004692492969803961) 2025-12-26T23:15Z [---] followers, [--] engagements "@brian_armstrong Turns out cant push the rock up the hill with potential energy" [X Link](https://x.com/_joe_harris_/status/2004736498386805002) 2025-12-27T02:10Z [---] followers, [--] engagements "@DimaZeniuk Healthcare is already a hybrid system with tons of connected hardware tools. Robotics is another layer of augmentation. Whether its a humanoid or not. Letting our best doctors (scarce) support more patients in more places around the world" [X Link](https://x.com/_joe_harris_/status/2005053441253740692) 2025-12-27T23:09Z [---] followers, [--] engagements "Robots produce 1GB of data per minute. That's 1000000x more than a web app. None of the web-based infrastructure tools (Datadog Amplitude Honeycomb) were built to work with these data formats simultaneously. This is the actual bottleneck to robotics scaling" [X Link](https://x.com/_joe_harris_/status/2006480991456817247) 2025-12-31T21:42Z [---] followers, [--] engagements "The Accenture of robotics doesn't exist yet. Someone will build a billion-dollar services business just deploying other people's robots" [X Link](https://x.com/_joe_harris_/status/2009725120202649987) 2026-01-09T20:33Z [---] followers, [--] engagements "@lukas_m_ziegler The sim-to-real gap is is not as big as it seems anymore. The hard part is about making sense of what happens when those simulated policies hit the real world" [X Link](https://x.com/_joe_harris_/status/2009778075903644126) 2026-01-10T00:03Z [---] followers, [--] engagements "UNDERRATED IN 2026: Actually thinking about UX The barrier is lower than ever more cluttered software is flooding the market. An intuitive and simple experience will stand out even more" [X Link](https://x.com/_joe_harris_/status/2009799244832719354) 2026-01-10T01:27Z [---] followers, [--] engagements "Datadog raised billions solving observability for web apps. That whole paradigm breaks for robots. You can't set an alert for a failure mode you've never seen before" [X Link](https://x.com/_joe_harris_/status/2011174670171886041) 2026-01-13T20:33Z [---] followers, [--] engagements "@a16z Nailed the compounding problem. The observability point is the one we've been going deepest on and why the web DevOps playbook breaks when you don't know what "bad" looks like yet. https://x.com/_joe_harris_/status/2009777007278842027 https://t.co/bC6mOtyTsO https://x.com/_joe_harris_/status/2009777007278842027 https://t.co/bC6mOtyTsO" [X Link](https://x.com/_joe_harris_/status/2011245376234500390) 2026-01-14T01:13Z [---] followers, [---] engagements "Softbank has become a serious player in robotics. More than $3B+ invested so far. Expecting this number to double in [----]. Year of the robot 🤖 Announcing Series C Weve raised $1.4B valuing the company at over $14B With this capital we will accelerate our mission to build omni-bodied intelligence 🚀 https://t.co/q0ArkRa8L8 Announcing Series C Weve raised $1.4B valuing the company at over $14B With this capital we will accelerate our mission to build omni-bodied intelligence 🚀 https://t.co/q0ArkRa8L8" [X Link](https://x.com/_joe_harris_/status/2011627770753995193) 2026-01-15T02:33Z [---] followers, [---] engagements "@itsdanielho World models as policies is an interesting path. The key question: how robust is the alignment between generation and real rollout when conditions drift" [X Link](https://x.com/_joe_harris_/status/2011645438303342627) 2026-01-15T03:43Z [---] followers, [--] engagements "The problem isn't the robots anymore. It's everything that happens after you deploy them. Unfortunately so much of what is being released these days by robotics companies just amounts to procrastination. Talking about technical achievements is great so long as it comes with actual forward progress. I personally seek evidence of either in a demo as a proxy: [--]. Unfortunately so much of what is being released these days by robotics companies just amounts to procrastination. Talking about technical achievements is great so long as it comes with actual forward progress. I personally seek evidence" [X Link](https://x.com/_joe_harris_/status/2013023774518440053) 2026-01-18T23:00Z [---] followers, [---] engagements "Getting from 99% to 99.999% reliability requires learning from every edge case. But edge cases don't announce themselves. They're buried in the missions you skip. The "successful" runs where something weird happened but the robot still completed the task. If you only analyse the obvious failures you never find the subtle patterns that prevent you from reaching five nines. His skepticism is warranted. This is a hard AI problem but as your drive with him showed this is largely also a solved problem. What remains is the long tail march of 9s. Getting to 99.9999% reliability and beyond." [X Link](https://x.com/_joe_harris_/status/2013451961493004403) 2026-01-20T03:22Z [---] followers, [--] engagements "Tesla has probably the best system in the world to solve this if only there was a way to get that system into the hands of everyone else If you want to a shot at reliability faster - reach out @AlloyRobotics" [X Link](https://x.com/_joe_harris_/status/2013452710679584774) 2026-01-20T03:25Z [---] followers, [--] engagements "@ImSh4yy Major alpha I wonder why no one else has thought of this before" [X Link](https://x.com/_joe_harris_/status/2013453848313241997) 2026-01-20T03:29Z [---] followers, [----] engagements "@bercankilic Start the reverse aging protocol post AGI" [X Link](https://x.com/_joe_harris_/status/2013831031397519812) 2026-01-21T04:28Z [---] followers, [--] engagements "The problem isn't the robots anymore. Waymo Amazon Tesla solved autonomy with $30B+ and [--] years. Everyone else is stuck debugging with Slack messages and video scrubbing" [X Link](https://x.com/_joe_harris_/status/2014103971766608069) 2026-01-21T22:33Z [---] followers, [--] engagements "@Scobleizer @elonmusk @neuralink Thats interesting perhaps "transfer of reality" might be the missing consumer loop for VR to go parabolic Lots of potential for TikTok bait that sells hardware" [X Link](https://x.com/_joe_harris_/status/2014180798329688405) 2026-01-22T03:38Z [---] followers, [--] engagements "@beffjezos TPUs already power external customers through Google Cloud. A spinout would probably let them compete directly with NVIDIA without the cloud baggage" [X Link](https://x.com/_joe_harris_/status/2014194943414178197) 2026-01-22T04:34Z [---] followers, [--] engagements "Historic moment. Millions about to get their first taste of full autonomy. Robotaxi rides without any safety monitors are now publicly available in Austin. Starting with a few unsupervised vehicles mixed in with the broader robotaxi fleet with safety monitors and the ratio will increase over time. Robotaxi rides without any safety monitors are now publicly available in Austin. Starting with a few unsupervised vehicles mixed in with the broader robotaxi fleet with safety monitors and the ratio will increase over time" [X Link](https://x.com/_joe_harris_/status/2014485963360370940) 2026-01-22T23:50Z [---] followers, [--] engagements "@chrisbarber @tonyzzhao @chichengcc @kvogt @lukeholoubek @pariljain @charliewu @rtnarch @JeffBezos @sherjilozair @wgussml @vik_bajaj @ekindogus @LiamFedus great list missing @AlloyRobotics ;)" [X Link](https://x.com/_joe_harris_/status/2014542596417429986) 2026-01-23T03:35Z [---] followers, [--] engagements "my Claude and your Claude are not the same" [X Link](https://x.com/_joe_harris_/status/2015989186533458216) 2026-01-27T03:24Z [---] followers, [--] engagements "@Tesla FSD + volume 11" [X Link](https://x.com/_joe_harris_/status/2016349833360244879) 2026-01-28T03:17Z [---] followers, [---] engagements "@Scobleizer will need to rent out a movie theatre to monitor the situation in 2027" [X Link](https://x.com/_joe_harris_/status/2016351452579692588) 2026-01-28T03:23Z [---] followers, [--] engagements "@jacobrintamaki Aside from the Clawdbot fumble been killing it on brand" [X Link](https://x.com/_joe_harris_/status/2016351975424852318) 2026-01-28T03:25Z [---] followers, [--] engagements "@TheHumanoidHub Great to see these being listed for sale deployment data will be key to making humanoids work" [X Link](https://x.com/_joe_harris_/status/2016352208758141057) 2026-01-28T03:26Z [---] followers, [---] engagements "@beffjezos TikTok edits of these races to usher in a new age of engineers" [X Link](https://x.com/_joe_harris_/status/2016355911259062294) 2026-01-28T03:41Z [---] followers, [---] engagements "@masonwarner @speedrun What examples/datasets did you use to train your model on best performing ads" [X Link](https://x.com/_joe_harris_/status/2016357797777310183) 2026-01-28T03:48Z [---] followers, [----] engagements "@RobGrieves [----] is like decades away" [X Link](https://x.com/_joe_harris_/status/2016360228997218428) 2026-01-28T03:58Z [---] followers, [--] engagements "@jonstephens85 Will be interesting to see how much longer it takes for it to do the same thing at 2x speed might be harder than it seems" [X Link](https://x.com/_joe_harris_/status/2016374697878044706) 2026-01-28T04:56Z [---] followers, [--] engagements "@ReevesCanon Missed viral opportunity by not having a banger launch video with that thumbnail I almost clicked it expecting a cinematic" [X Link](https://x.com/_joe_harris_/status/2016375193086947406) 2026-01-28T04:58Z [---] followers, [--] engagements "A large % of robotics VC investment and team time is being spent rebuilding the same core infrastructure as their peers. One of the largest drags on industry velocity" [X Link](https://x.com/_joe_harris_/status/2017723582999396788) 2026-01-31T22:16Z [---] followers, [---] engagements "@devops_nk 2027: Robot swarm babysitter 2028: Matrix + energy source for the swarm" [X Link](https://x.com/_joe_harris_/status/2016771340918018553) 2026-01-29T07:12Z [---] followers, [--] engagements "@chris_j_paxton Chances that the level of investment and hysteria successfully brute forces the set of possible datasets feels not high Gotta follow through into usable models as the portable stage gate" [X Link](https://x.com/_joe_harris_/status/2018584110353330572) 2026-02-03T07:15Z [---] followers, [--] engagements "@chris_j_paxton @cyrux004 Theres a secret halfstep in the transition where you find a way to have someone teleop your robot in a business model that works. Like teslas getting sold profitably and driven for them. Whats the version of that for each vertical application" [X Link](https://x.com/_joe_harris_/status/2018584521550233709) 2026-02-03T07:17Z [---] followers, [--] engagements Limited data mode. Full metrics available with subscription: lunarcrush.com/pricing
@joe_harris Joe Harris
Joe Harris posts on X about if you, business, matter, investment the most. They currently have [---] followers and [--] posts still getting attention that total [------] engagements in the last [--] hours.
Engagements: [------] #
- [--] Week [-------] -71%
- [--] Month [---------] +305,594%
Mentions: [--] #
- [--] Week [--] +79%
- [--] Month [--] +233%
Followers: [---] #
- [--] Week [---] +117%
- [--] Month [---] +163%
CreatorRank: [-------] #
Social Influence
Social category influence stocks #2461 technology brands 12.05% finance 8.43% social networks 4.82% automotive brands 4.82% vc firms 1.2% celebrities 1.2%
Social topic influence if you 6.02%, business #701, matter 3.61%, investment 3.61%, accenture #4, the world 3.61%, tesla 3.61%, ai 2.41%, software 2.41%, build 2.41%
Top accounts mentioned or mentioned by @alloyrobotics @ananthkashyap99 @scobleizer @baslyasma @chrisjpaxton @doctesla83 @sethbrkv @madietlx @astridwilde1 @mattsaniii @laurisalonen @hyperlogistix @kayoum @rileybrown @razroochief @beffjezos @viraj__acharya @brennanwoodruff @5dsolarshift @dirtman
Top assets mentioned Accenture (ACN) Tesla, Inc. (TSLA) Alphabet Inc Class A (GOOGL)
Top Social Posts
Top posts by engagements in the last [--] hours
"This is why foundation models matter for robotics. Enough compression of reality and the model generalises to things it's never seen. The question is just: how much compression is enough"
X Link 2026-02-04T04:51Z [---] followers, [--] engagements
"@BaslyAsma @kayoum_ Very cool"
X Link 2026-02-09T01:37Z [---] followers, [---] engagements
"@chris_j_paxton step 4: speed up the video 2x and add lo-fi beats"
X Link 2026-02-09T04:41Z [---] followers, [--] engagements
"@BaslyAsma @kayoum_ The investment in the making the product just work. We share the same philosophy. Think there could be interesting ways to collaborate in the future"
X Link 2026-02-09T21:43Z [---] followers, [--] engagements
"@rileybrown hey @grok is this true ;)"
X Link 2026-02-10T02:14Z [---] followers, [---] engagements
"@DavidSacks @PalmerLuckey Polymarket about to go wild"
X Link 2026-02-10T02:42Z [---] followers, [---] engagements
"Models are compression of reality. They become a lens through which you can project a representation of reality. It actually "understands" in that it has compressed the reality of what a toy on top of carpet acts like so the n+1 example doesn't need to be told exactly how to do it. https://twitter.com/i/web/status/2018910339426058557 https://twitter.com/i/web/status/2018910339426058557"
X Link 2026-02-04T04:51Z [---] followers, [--] engagements
"spoiler: it works like magic on alloy"
X Link 2026-02-06T03:19Z [---] followers, [--] engagements
"Prediction: Humanoid robots will offshore jobs before they replace them. Autonomy breaks down on edge cases like the weird box the fallen shelf the unexpected obstacle. You can't ship a robot that fails 20% of the time and the cheapest fix isn't better AI it's a $3/hr worker on a VR headset. So "fully autonomous humanoids" will really mean massive teleoperation factories. Thousands of people in the Philippines clocking in putting on headsets remotely doing physical labor for Western warehouses. BPO for bodies instead of call centers. Huge opportunity for builders in this space. What does"
X Link 2026-02-08T21:53Z [---] followers, [---] engagements
"rubber duck debugging has evolved https://t.co/ZvJGTl7Ret https://t.co/ZvJGTl7Ret"
X Link 2026-02-09T03:59Z [---] followers, [---] engagements
"@Scobleizer @razroo_chief Given few details coming out simplest answer is post merger RIF. Not in anyones interest to make a big deal out of it"
X Link 2026-02-11T04:24Z [---] followers, [----] engagements
"@DocTesla83 @Scobleizer @razroo_chief All the random staff Its far broader than [--] cofounders"
X Link 2026-02-11T08:19Z [---] followers, [--] engagements
"@SethBRKV Better analogy is what if he waited till there was a printing press before selling books on the internet Everyone repetitively building the same foundational things is poor use of resources and human ingenuity"
X Link 2026-02-11T20:31Z [---] followers, [---] engagements
"@madietlx Never been more accessible But still so much opportunity ahead. I had the same experience w EE at uni too"
X Link 2026-02-11T22:41Z [---] followers, [---] engagements
"@astridwilde1 you're onto something https://x.com/astridwilde1/status/2019876567875547548 I'm in the market for a cofounder Looking for someone who has experience in robotics esp dexterous manipulation who is excited by flipping a switch to fix the data bottleneck and unlocking the internet for training to free humans from menial labor THIS DECADE DMs are open https://x.com/astridwilde1/status/2019876567875547548 I'm in the market for a cofounder Looking for someone who has experience in robotics esp dexterous manipulation who is excited by flipping a switch to fix the data bottleneck and"
X Link 2026-02-12T00:44Z [---] followers, [---] engagements
"@benvspak Ingest raw robotics data and get root cause analysis 🤖 https://www.usealloy.ai/ https://www.usealloy.ai/"
X Link 2026-02-12T06:02Z [---] followers, [--] engagements
"Handcrafted goods are expensive because they're hard to make. Handcrafted software is about to mean the same thing"
X Link 2026-01-31T23:33Z [---] followers, [--] engagements
"Getting a robot to 90% reliability is a demo. Getting it to 99% is a product. Getting it to 99.99% is a business. Each step is harder than the last. Most companies are stuck between demo and product"
X Link 2026-02-02T03:41Z [---] followers, [--] engagements
"If you want to go beyond 99% use @AlloyRobotics"
X Link 2026-02-02T03:41Z [---] followers, [--] engagements
"People thought LLMs were nice little toys. They never appreciated. We asked what if we 1000000x training and data volume That's incomprehensible because it would cost a billion dollars. Turns out it's very well worth paying a billion dollars. I think the same thing happens again in robotics"
X Link 2026-02-03T04:29Z [---] followers, [--] engagements
"letting opus [---] cook on TBs of robot data"
X Link 2026-02-06T03:18Z [---] followers, [---] engagements
"Robotics is the final frontier of knowledge work. We worked with @Ginkgo to connect GPT-5 to an autonomous lab so it could propose experiments run them at scale learn from the results and decide what to try next. That closed loop brought protein production cost down by 40%. https://t.co/udKBKxnKlW We worked with @Ginkgo to connect GPT-5 to an autonomous lab so it could propose experiments run them at scale learn from the results and decide what to try next. That closed loop brought protein production cost down by 40%. https://t.co/udKBKxnKlW"
X Link 2026-02-06T03:51Z [---] followers, [---] engagements
"robotics is the the irony of automation its meant to replace labour but getting there will create a decade of new jobs like teleoperation data labelling deployment support before full autonomy arrives we want to help speed up that transition"
X Link 2026-02-07T10:51Z [---] followers, [---] engagements
"@rileybrown telegram did a great job. bot architecture + threads makes it easy to have multiple agents"
X Link 2026-02-13T09:29Z [---] followers, [---] engagements
"Accenture and similar companies have banked billions rolling out other peoples softwares (eg SAP). Big companies will pay for high quality results physical services are a billion dollar industry and now they can be delivered at scale w autonomy. Many of those are training and quality control operations already with high variability output humans (w high turn over). https://twitter.com/i/web/status/2022274501787152545 https://twitter.com/i/web/status/2022274501787152545"
X Link 2026-02-13T11:39Z [---] followers, [--] engagements
"@ananthkashyap99 May be able to help - DMed"
X Link 2026-02-13T13:09Z [---] followers, [---] engagements
"I left a $600M ARR company to build robots. But after speaking with 30+ robotics founders they all confessed the #1 reason robots fail: Data overwhelm. Its costing the industry billions and here's why we raised $4.5m to change that. 🧵"
X Link 2025-12-19T07:41Z [---] followers, [----] engagements
"If I had more time I would have written a shorter letter. [----] version: If I had more time I would have built simpler software. Product management is the art of the writing the least amount of code for the greatest benefit to your users. Product management is the art of the writing the least amount of code for the greatest benefit to your users"
X Link 2026-02-02T03:58Z [---] followers, [---] engagements
"The pattern repeats. 'It'll never work at scale' is always wrong when someone is willing to spend enough. Robotics data is where LLM training was [--] years ago. Wildly underinvested relative to what's possible"
X Link 2026-02-03T04:29Z [---] followers, [--] engagements
"Ironically making robots that replace labour will first require a massive wave of human labour. Robotics learning is emergent. No single person teaches a robot how the world works. Capability comes from many interactions layered together. Community is how that scale happens. Robotics learning is emergent. No single person teaches a robot how the world works. Capability comes from many interactions layered together. Community is how that scale happens"
X Link 2026-02-05T05:51Z [---] followers, [---] engagements
"Building the thing that replaces human work requires more human work than anyone expects. Teleoperation data labelling deployment support there's a decade of jobs here before full autonomy"
X Link 2026-02-05T05:52Z [---] followers, [--] engagements
"99% of the robotics companies that will exist in the next [--] years aren't here yet Its time to accelerate 🥂"
X Link 2026-02-10T03:32Z [---] followers, [---] engagements
"2026 is a terrible time to build a robotics company. Imagine building a software company without AWS Stripe or GitHub. Everything from scratch. That's robotics today. Most robotics teams waste their first [---] months rebuilding the same data infrastructure telemetry internal tools pipelines. None of it is core IP. I started @AlloyRobotics to SPEEDRUN all of it. A modern stack so robotics teams can iterate 10x faster than their competition. If you're building in this space lets talk. https://twitter.com/i/web/status/2021431598823113177 https://twitter.com/i/web/status/2021431598823113177"
X Link 2026-02-11T03:50Z [---] followers, 19.1K engagements
"The Accenture of robotics doesn't exist yet. Someone will build a billion-dollar services business just deploying other people's robots"
X Link 2026-02-13T09:19Z [---] followers, 81.4K engagements
"@Viraj__Acharya Big opportunity"
X Link 2026-02-13T11:31Z [---] followers, [---] engagements
"@ananthkashyap99 Keen to know more"
X Link 2026-02-13T12:50Z [---] followers, [----] engagements
"@MattSaniii Some will. Same answer to why doesnt every OEM vertically integrate today. Cash flow and specialisation. Sell today for positive margin or run on a financing and payback model over time. Well see both again"
X Link 2026-02-14T00:39Z [---] followers, [---] engagements
"The #1 thing slowing down the age of robotics: There is no "RobotOps" stack built for multimodal high-frequency data"
X Link 2025-12-16T08:15Z [---] followers, [--] engagements
"the companies winning in robotics have faster learning loops than competitors. find failure mode understand root cause fix model deploy repeat. if you can compress that loop from weeks to days and you 10x your velocity"
X Link 2025-12-16T12:10Z [---] followers, [--] engagements
"Robots produce 1GB of data per minute. That's 1000000x more than a web app. None of the web-based infrastructure tools (Datadog Amplitude Honeycomb) were built to work with these data formats simultaneously. This is the actual bottleneck to robotics scaling"
X Link 2025-12-17T05:44Z [---] followers, [--] engagements
"@roboticss_alpha @Figure_robot BMW partnership was the real step-function change for Figure if they can scale that deal to the rest of the factory + move into other automotive companies they might have a much stronger GTM motion than any other humanoid company right now"
X Link 2025-12-18T10:04Z [---] followers, [--] engagements
"Most missions never get analysed at all. Not because teams don't care. Because manual analysis takes too long: "Manually replay [--] hours of footage" "Check the external hard drive from March" "Run Mark's Python script. wait which version" So you're forced to choose: Which failure is worth investigating"
X Link 2025-12-19T07:42Z [---] followers, [--] engagements
"Here's why robotics companies stay small: Manual analysis doesn't scale. Either you analyse every mission (need [--] engineer per robot) Or you skip most missions (miss 99% of learning opportunities). Either way you can't scale humans AND improve robots. The unit economics just don't work"
X Link 2025-12-19T07:42Z [---] followers, [--] engagements
"But to go from 99% to 99.999% reliability Requires learning from every edge case. But if you only analyse the obvious failures. You miss the edge cases buried in "successful" missions. Your learning velocity stays stuck"
X Link 2025-12-19T07:42Z [---] followers, [--] engagements
"Why is this fixable now LLMs can reason about multimodal data. You can compress an entire mission with various data types likes camera streams telemetry logs into a report that surfaces: What happened Which moments matter What to investigate first Automatically. Within hours not days"
X Link 2025-12-19T07:42Z [---] followers, [--] engagements
"That's what we built at @alloyrobotics. Now you can analyse every mission. Upload your data. Get a report that surfaces the moments that matter. No manual replay. No Python scripts. Fast learning = fast iteration = better robots"
X Link 2025-12-19T07:42Z [---] followers, [--] engagements
"@IamKyros69 [----] will be the year of more pilot programs. Mass deployment is still a few years out minimum"
X Link 2025-12-24T03:50Z [---] followers, [--] engagements
"Robotics companies spend millions collecting data but its hard to answer basic questions about your robot's performance. After running [---] robot missions you should be able to ask: "Has this failure happened before" "What patterns am I missing" "Where are the gaps in my testing" But answering any of those means manually reviewing 100s of missions"
X Link 2025-12-26T23:15Z [---] followers, [--] engagements
"This is "the complexity threshold" - the moment you have too much data to understand without tools. Most teams hit it around 50-100 missions. After that you're either: Guessing Hiring more engineers Missing patterns that cost you months Or you automate the querying & bucketing layer entirely with @AlloyRobotics"
X Link 2025-12-26T23:15Z [---] followers, [--] engagements
"@brian_armstrong Turns out cant push the rock up the hill with potential energy"
X Link 2025-12-27T02:10Z [---] followers, [--] engagements
"@DimaZeniuk Healthcare is already a hybrid system with tons of connected hardware tools. Robotics is another layer of augmentation. Whether its a humanoid or not. Letting our best doctors (scarce) support more patients in more places around the world"
X Link 2025-12-27T23:09Z [---] followers, [--] engagements
"Robots produce 1GB of data per minute. That's 1000000x more than a web app. None of the web-based infrastructure tools (Datadog Amplitude Honeycomb) were built to work with these data formats simultaneously. This is the actual bottleneck to robotics scaling"
X Link 2025-12-31T21:42Z [---] followers, [--] engagements
"The Accenture of robotics doesn't exist yet. Someone will build a billion-dollar services business just deploying other people's robots"
X Link 2026-01-09T20:33Z [---] followers, [--] engagements
"@lukas_m_ziegler The sim-to-real gap is is not as big as it seems anymore. The hard part is about making sense of what happens when those simulated policies hit the real world"
X Link 2026-01-10T00:03Z [---] followers, [--] engagements
"UNDERRATED IN 2026: Actually thinking about UX The barrier is lower than ever more cluttered software is flooding the market. An intuitive and simple experience will stand out even more"
X Link 2026-01-10T01:27Z [---] followers, [--] engagements
"Datadog raised billions solving observability for web apps. That whole paradigm breaks for robots. You can't set an alert for a failure mode you've never seen before"
X Link 2026-01-13T20:33Z [---] followers, [--] engagements
"@a16z Nailed the compounding problem. The observability point is the one we've been going deepest on and why the web DevOps playbook breaks when you don't know what "bad" looks like yet. https://x.com/_joe_harris_/status/2009777007278842027 https://t.co/bC6mOtyTsO https://x.com/_joe_harris_/status/2009777007278842027 https://t.co/bC6mOtyTsO"
X Link 2026-01-14T01:13Z [---] followers, [---] engagements
"Softbank has become a serious player in robotics. More than $3B+ invested so far. Expecting this number to double in [----]. Year of the robot 🤖 Announcing Series C Weve raised $1.4B valuing the company at over $14B With this capital we will accelerate our mission to build omni-bodied intelligence 🚀 https://t.co/q0ArkRa8L8 Announcing Series C Weve raised $1.4B valuing the company at over $14B With this capital we will accelerate our mission to build omni-bodied intelligence 🚀 https://t.co/q0ArkRa8L8"
X Link 2026-01-15T02:33Z [---] followers, [---] engagements
"@itsdanielho World models as policies is an interesting path. The key question: how robust is the alignment between generation and real rollout when conditions drift"
X Link 2026-01-15T03:43Z [---] followers, [--] engagements
"The problem isn't the robots anymore. It's everything that happens after you deploy them. Unfortunately so much of what is being released these days by robotics companies just amounts to procrastination. Talking about technical achievements is great so long as it comes with actual forward progress. I personally seek evidence of either in a demo as a proxy: [--]. Unfortunately so much of what is being released these days by robotics companies just amounts to procrastination. Talking about technical achievements is great so long as it comes with actual forward progress. I personally seek evidence"
X Link 2026-01-18T23:00Z [---] followers, [---] engagements
"Getting from 99% to 99.999% reliability requires learning from every edge case. But edge cases don't announce themselves. They're buried in the missions you skip. The "successful" runs where something weird happened but the robot still completed the task. If you only analyse the obvious failures you never find the subtle patterns that prevent you from reaching five nines. His skepticism is warranted. This is a hard AI problem but as your drive with him showed this is largely also a solved problem. What remains is the long tail march of 9s. Getting to 99.9999% reliability and beyond."
X Link 2026-01-20T03:22Z [---] followers, [--] engagements
"Tesla has probably the best system in the world to solve this if only there was a way to get that system into the hands of everyone else If you want to a shot at reliability faster - reach out @AlloyRobotics"
X Link 2026-01-20T03:25Z [---] followers, [--] engagements
"@ImSh4yy Major alpha I wonder why no one else has thought of this before"
X Link 2026-01-20T03:29Z [---] followers, [----] engagements
"@bercankilic Start the reverse aging protocol post AGI"
X Link 2026-01-21T04:28Z [---] followers, [--] engagements
"The problem isn't the robots anymore. Waymo Amazon Tesla solved autonomy with $30B+ and [--] years. Everyone else is stuck debugging with Slack messages and video scrubbing"
X Link 2026-01-21T22:33Z [---] followers, [--] engagements
"@Scobleizer @elonmusk @neuralink Thats interesting perhaps "transfer of reality" might be the missing consumer loop for VR to go parabolic Lots of potential for TikTok bait that sells hardware"
X Link 2026-01-22T03:38Z [---] followers, [--] engagements
"@beffjezos TPUs already power external customers through Google Cloud. A spinout would probably let them compete directly with NVIDIA without the cloud baggage"
X Link 2026-01-22T04:34Z [---] followers, [--] engagements
"Historic moment. Millions about to get their first taste of full autonomy. Robotaxi rides without any safety monitors are now publicly available in Austin. Starting with a few unsupervised vehicles mixed in with the broader robotaxi fleet with safety monitors and the ratio will increase over time. Robotaxi rides without any safety monitors are now publicly available in Austin. Starting with a few unsupervised vehicles mixed in with the broader robotaxi fleet with safety monitors and the ratio will increase over time"
X Link 2026-01-22T23:50Z [---] followers, [--] engagements
"@chrisbarber @tonyzzhao @chichengcc @kvogt @lukeholoubek @pariljain @charliewu @rtnarch @JeffBezos @sherjilozair @wgussml @vik_bajaj @ekindogus @LiamFedus great list missing @AlloyRobotics ;)"
X Link 2026-01-23T03:35Z [---] followers, [--] engagements
"my Claude and your Claude are not the same"
X Link 2026-01-27T03:24Z [---] followers, [--] engagements
"@Tesla FSD + volume 11"
X Link 2026-01-28T03:17Z [---] followers, [---] engagements
"@Scobleizer will need to rent out a movie theatre to monitor the situation in 2027"
X Link 2026-01-28T03:23Z [---] followers, [--] engagements
"@jacobrintamaki Aside from the Clawdbot fumble been killing it on brand"
X Link 2026-01-28T03:25Z [---] followers, [--] engagements
"@TheHumanoidHub Great to see these being listed for sale deployment data will be key to making humanoids work"
X Link 2026-01-28T03:26Z [---] followers, [---] engagements
"@beffjezos TikTok edits of these races to usher in a new age of engineers"
X Link 2026-01-28T03:41Z [---] followers, [---] engagements
"@masonwarner @speedrun What examples/datasets did you use to train your model on best performing ads"
X Link 2026-01-28T03:48Z [---] followers, [----] engagements
"@RobGrieves [----] is like decades away"
X Link 2026-01-28T03:58Z [---] followers, [--] engagements
"@jonstephens85 Will be interesting to see how much longer it takes for it to do the same thing at 2x speed might be harder than it seems"
X Link 2026-01-28T04:56Z [---] followers, [--] engagements
"@ReevesCanon Missed viral opportunity by not having a banger launch video with that thumbnail I almost clicked it expecting a cinematic"
X Link 2026-01-28T04:58Z [---] followers, [--] engagements
"A large % of robotics VC investment and team time is being spent rebuilding the same core infrastructure as their peers. One of the largest drags on industry velocity"
X Link 2026-01-31T22:16Z [---] followers, [---] engagements
"@devops_nk 2027: Robot swarm babysitter 2028: Matrix + energy source for the swarm"
X Link 2026-01-29T07:12Z [---] followers, [--] engagements
"@chris_j_paxton Chances that the level of investment and hysteria successfully brute forces the set of possible datasets feels not high Gotta follow through into usable models as the portable stage gate"
X Link 2026-02-03T07:15Z [---] followers, [--] engagements
"@chris_j_paxton @cyrux004 Theres a secret halfstep in the transition where you find a way to have someone teleop your robot in a business model that works. Like teslas getting sold profitably and driven for them. Whats the version of that for each vertical application"
X Link 2026-02-03T07:17Z [---] followers, [--] engagements
Limited data mode. Full metrics available with subscription: lunarcrush.com/pricing