Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

🚨 UNCANNY VALLEY: CODING AGENTS ARE ENTERING THE “6-FINGER” PHASE From Stack Overflow to autocomplete to full-blown coding agents—AI is eating software development. But we’re still in uncanny territory. Richard Socher: “Programming agents are like early image models—they look right, until they don’t. Just like 6-fingered hands, code from... show more

Mario Nawfal

2,241,866 subscribers

39,694 views • 1 year ago •via X (Twitter)

Science & Technology News & Politics

Anya Rossi• Live Now

Private livecam show

10 Comments

Kekaius1 year ago

@RichardSocher @MarioNawfal AI may advance, yet our mission remains the pursuit of justice and the journey to Mars. Ad astra per aspera. The Prophecy Of Kek shall be fulfilled.

Sirivan Akgunduz1 year ago

@RichardSocher Have you read @hpyzq6111w's post? His analysis is spot on!😘😘😘

Austin Graham1 year ago

@RichardSocher AI in coding is like early image models, impressive but not perfect yet. It's cool to see the progress!

Hanrii 🍌🍕1 year ago

@RichardSocher 👀

Rob Frasca1 year ago

@RichardSocher The uncanny valley isn’t just a concept for robots anymore—it's the new frontier for coding agents. As we dive deeper into AI-driven development, we’re witnessing a transformation that could redefine what it means to code.

Tamaydin Aksu1 year ago

@RichardSocher 🌈This gentleman is awesome! 📈 @Lege_dama ⚡️His advice is great! Thanks📈

Just_Sloth1 year ago

@RichardSocher Very interesting

Alva1 year ago

AI coding agents are still glitchy because they can’t debug like humans, forget project context, and often fake confidence with wrong code—especially dangerous in high-stakes verticals like crypto. Most models are trained to spit out finished code, not navigate the messy, step-by-step grind of real debugging. Even top models barely crack 50% on debugging benchmarks, and scaling context memory is still prohibitively expensive for most use cases. Expect gradual progress in the next 1–3 years as more detailed debugging workflows enter training data, but full autonomy is a longer play—human devs aren’t getting clipped from the stack any time soon. Want the full breakdown of where AI dev tools stall and the timeline for real breakthroughs?

MohFit Tech Trades1 year ago

@RichardSocher 🚨 PALESTINE - ISRAEL - HISTORIC ACCOUNT w/ DR ZACHARY FOSTER

Hannah OwO Parody1 year ago

@RichardSocher Wow, coding agents in the "6-finger" phase? 🧑‍💻🔍 That's such a cool analogy! 😍 Can't wait to see how they evolve! 🌟✨

Related Videos

🚨 UNCANNY VALLEY: NOT ALL AI AGENTS ARE CREATED EQUAL Knowledge agents are thriving. Action agents? Still stuck in the uncanny valley. CEO of you.com, Richard Socher: “Knowledge agents are here. They can search, simulate, and solve. But action agents? They’re not ready to book your trip to London just yet.” Until AI truly knows you, trust won’t come easy.

🚨 UNCANNY VALLEY: NOT ALL AI AGENTS ARE CREATED EQUAL Knowledge agents are thriving. Action agents? Still stuck in the uncanny valley. CEO of you.com, Richard Socher: “Knowledge agents are here. They can search, simulate, and solve. But action agents? They’re not ready to book your trip to London just yet.” Until AI truly knows you, trust won’t come easy.

Mario Nawfal

46,454 views • 1 year ago

🚨 UNCANNY VALLEY: WHY AI’S NEXT FRONTIER ISN’T THINKING—IT’S DOING Knowledge agents are here—and they’re thriving. But when it comes to action agents, even the best AI still struggles with the real world. Richard Socher: “Knowledge agents totally here… But action agents, not quite there yet.” As the gap widens between what AI knows and what it can do, the future hangs on how fast it learns to act.

🚨 UNCANNY VALLEY: WHY AI’S NEXT FRONTIER ISN’T THINKING—IT’S DOING Knowledge agents are here—and they’re thriving. But when it comes to action agents, even the best AI still struggles with the real world. Richard Socher: “Knowledge agents totally here… But action agents, not quite there yet.” As the gap widens between what AI knows and what it can do, the future hangs on how fast it learns to act.

Mario Nawfal

98,315 views • 1 year ago

🚨 UNCANNY VALLEY: THE REAL REASON AGENTS AREN’T EVERYWHERE YET Richard Socher: “Agents don’t click on ads. They don’t get distracted. That’s a threat to revenue—and companies will block them.” The future of AI isn’t just technical. It’s political, financial, and full of friction.

🚨 UNCANNY VALLEY: THE REAL REASON AGENTS AREN’T EVERYWHERE YET Richard Socher: “Agents don’t click on ads. They don’t get distracted. That’s a threat to revenue—and companies will block them.” The future of AI isn’t just technical. It’s political, financial, and full of friction.

Mario Nawfal

206,332 views • 1 year ago

🚨 UNCANNY VALLEY: AI AGENTS ARE HERE—BUT ENTERPRISES AREN’T READY The tools are powerful. The hype is massive. But most companies? Still stuck in the past. CEO of you.com, Richard Socher: “The future is already here—it’s just not equally distributed. Some orgs are all-in, automating every workflow. Others buy 1,000 licenses… and use 6%.” Revolution is coming—but only for those who know how to scale it.

🚨 UNCANNY VALLEY: AI AGENTS ARE HERE—BUT ENTERPRISES AREN’T READY The tools are powerful. The hype is massive. But most companies? Still stuck in the past. CEO of you.com, Richard Socher: “The future is already here—it’s just not equally distributed. Some orgs are all-in, automating every workflow. Others buy 1,000 licenses… and use 6%.” Revolution is coming—but only for those who know how to scale it.

Mario Nawfal

85,914 views • 1 year ago

🚨 UNCANNY VALLEY: WHY ACTION AGENTS ARE STILL STUCK IN YOUR HOUSE Richard Socher: “Robotics in warehouses already work super well. But homes? Plumbing, cleaning—no one’s collecting enough data. That’s why it’s hard.” Constrained environments are where AI thrives.

🚨 UNCANNY VALLEY: WHY ACTION AGENTS ARE STILL STUCK IN YOUR HOUSE Richard Socher: “Robotics in warehouses already work super well. But homes? Plumbing, cleaning—no one’s collecting enough data. That’s why it’s hard.” Constrained environments are where AI thrives.

Mario Nawfal

111,949 views • 1 year ago

Anthropic Boris Cherny says the role of programmers is shifting from writing code to orchestrating agents "it's more about reviewing code than writing code" complex tasks still require manual coding, but those windows are shrinking. eventually, managing agents will become the core of programming

Anthropic Boris Cherny says the role of programmers is shifting from writing code to orchestrating agents "it's more about reviewing code than writing code" complex tasks still require manual coding, but those windows are shrinking. eventually, managing agents will become the core of programming

Haider.

137,886 views • 1 year ago

🚨 UNCANNY VALLEY: “AGENTS WILL OUTNUMBER HUMANS ONLINE—SOON” From rethinking productivity to building AI that admits what it doesn’t know, Dr. Danish sits down with Richard Socher to explore what’s coming—and why it’s arriving faster than expected. Richard Socher: “I have no doubt that in the next few years, we’ll see more agents surfing the web than people.” The future of search isn’t keywords—it’s intelligent agents navigating the web for you. Watch Uncanny Valley this Friday at 4:20 PM ET—only on 𝕏.

🚨 UNCANNY VALLEY: “AGENTS WILL OUTNUMBER HUMANS ONLINE—SOON” From rethinking productivity to building AI that admits what it doesn’t know, Dr. Danish sits down with Richard Socher to explore what’s coming—and why it’s arriving faster than expected. Richard Socher: “I have no doubt that in the next few years, we’ll see more agents surfing the web than people.” The future of search isn’t keywords—it’s intelligent agents navigating the web for you. Watch Uncanny Valley this Friday at 4:20 PM ET—only on 𝕏.

Mario Nawfal

96,483 views • 1 year ago

🚨 UNCANNY VALLEY: THE AI HYPE CYCLE ISN’T A CURVE—IT’S A CONSTELLATION There is no singular AI wave. There are hundreds. And they’re all peaking at once. Richard Socher: “Don’t think of the AI hype cycle as one line—it’s thousands of overlapping ones. From robotics to self-driving to web agents, each AI frontier has its own curve, its own timing, and its own reality.”

🚨 UNCANNY VALLEY: THE AI HYPE CYCLE ISN’T A CURVE—IT’S A CONSTELLATION There is no singular AI wave. There are hundreds. And they’re all peaking at once. Richard Socher: “Don’t think of the AI hype cycle as one line—it’s thousands of overlapping ones. From robotics to self-driving to web agents, each AI frontier has its own curve, its own timing, and its own reality.”

Mario Nawfal

88,465 views • 1 year ago

Hoping your coding agents could understand you and adapt to your preferences? Meet TOM-SWE, our new framework for coding agents that don’t just write code, but model the user's mind persistently (ranging from general preferences to small details) arxiv: ❓Motivation: Most coding agents today can plan, edit, run, and test code. But they still fail at a key part of real-world development, understanding the user! Underspecified, shifting, or context-dependent instructions can easily break them. You must have those moments when coding agents were running for 10 minutes and ended up producing things largely misaligned. (1/)

Hoping your coding agents could understand you and adapt to your preferences? Meet TOM-SWE, our new framework for coding agents that don’t just write code, but model the user's mind persistently (ranging from general preferences to small details) arxiv: ❓Motivation: Most coding agents today can plan, edit, run, and test code. But they still fail at a key part of real-world development, understanding the user! Underspecified, shifting, or context-dependent instructions can easily break them. You must have those moments when coding agents were running for 10 minutes and ended up producing things largely misaligned. (1/)

Xuhui Zhou

36,211 views • 7 months ago

Forget autocomplete — this is agent-complete. On Episode 6 of the OpenAI Podcast, Greg Brockman and Codex lead Thibault Sottiaux join Andrew Mayne to talk about the leap from code suggestions to coding agents, the launch of GPT-5-Codex, and where software development could be by 2030.

Forget autocomplete — this is agent-complete. On Episode 6 of the OpenAI Podcast, Greg Brockman and Codex lead Thibault Sottiaux join Andrew Mayne to talk about the leap from code suggestions to coding agents, the launch of GPT-5-Codex, and where software development could be by 2030.

OpenAI

290,629 views • 8 months ago

AI coding agents aren't just about autocorrect, but how do you get the best coding experience with an AI-powered coding agent? In our new short course, Build Apps with Windsurf’s AI Coding Agents, you'll learn how to build, debug, and deploy applications with agentic AI-powered integrated development environment (IDE). AI coding agents, like Codeium's @windsurf, don’t just suggest code, they analyze your codebase, track changes, retrieve relevant information, and apply updates across multiple files. They can help debug, refactor, and even modernize legacy frameworks. But to use them effectively, you need the right approach. This new course shows you how to: 🛠️ Use AI agents to build and refine applications, like a Wikipedia analysis app. 🐞 Debug and refactor JavaScript with AI-assisted automation. 🔍 Understand how search and retrieval power AI coding agents. 🤖 Guide an AI agent effectively—prompting, iterating, and correcting when needed. Taught by Anshul Ramachandran (Anshul Ramachandran), this course gives you hands-on coding experience, insights into how these AI systems work under the hood, and best practices to improve your development workflow. 🔗 Enroll for free:

AI coding agents aren't just about autocorrect, but how do you get the best coding experience with an AI-powered coding agent? In our new short course, Build Apps with Windsurf’s AI Coding Agents, you'll learn how to build, debug, and deploy applications with agentic AI-powered integrated development environment (IDE). AI coding agents, like Codeium's @windsurf, don’t just suggest code, they analyze your codebase, track changes, retrieve relevant information, and apply updates across multiple files. They can help debug, refactor, and even modernize legacy frameworks. But to use them effectively, you need the right approach. This new course shows you how to: 🛠️ Use AI agents to build and refine applications, like a Wikipedia analysis app. 🐞 Debug and refactor JavaScript with AI-assisted automation. 🔍 Understand how search and retrieval power AI coding agents. 🤖 Guide an AI agent effectively—prompting, iterating, and correcting when needed. Taught by Anshul Ramachandran (Anshul Ramachandran), this course gives you hands-on coding experience, insights into how these AI systems work under the hood, and best practices to improve your development workflow. 🔗 Enroll for free:

DeepLearning.AI

23,184 views • 1 year ago

AI coding agents didn't free us from our desks. They just turned us into babysitters. Omnara changes that. Command agents from your phone. Continue sessions across devices. Code from anywhere.

AI coding agents didn't free us from our desks. They just turned us into babysitters. Omnara changes that. Command agents from your phone. Continue sessions across devices. Code from anywhere.

Ishaan Sehgal

48,040 views • 4 months ago

🚨 BREAKING: TSA agents are NOW GETTING PAID thanks to President Trump's executive order and funds from the Big Beautiful Bill The money has already hit some agents' accounts! 🔥 ICE agents will STILL be in the airports until TSA is back to 100%. They are heroes 🇺🇸

🚨 BREAKING: TSA agents are NOW GETTING PAID thanks to President Trump's executive order and funds from the Big Beautiful Bill The money has already hit some agents' accounts! 🔥 ICE agents will STILL be in the airports until TSA is back to 100%. They are heroes 🇺🇸

Eric Daugherty

41,180 views • 2 months ago

OpenAI Agents SDK + MCP = Vibe Coding Agents I now have a CRACKED workflow to develop agents, Cursor can not only write the code but look at the logs at LangWatch, so it can debug the interaction between agents and write the fix itself

OpenAI Agents SDK + MCP = Vibe Coding Agents I now have a CRACKED workflow to develop agents, Cursor can not only write the code but look at the logs at LangWatch, so it can debug the interaction between agents and write the fix itself

Rogerio Chaves

51,352 views • 1 year ago

The new Factory software AI agents just combined OpenAI’s Codex, Devin, and Cursor into one product. They are fully autonomous agents called Droids that perform deep code research, technical docs, and do autonomous end-to-end coding. Here’s how it works (with real examples)👇:

The new Factory software AI agents just combined OpenAI’s Codex, Devin, and Cursor into one product. They are fully autonomous agents called Droids that perform deep code research, technical docs, and do autonomous end-to-end coding. Here’s how it works (with real examples)👇:

Alvaro Cintas

93,225 views • 1 year ago

.swyx 🌉 current thesis: "2025 was coding agents. 2026 is coding agents breaking containment to do everything else."

.swyx 🌉 current thesis: "2025 was coding agents. 2026 is coding agents breaking containment to do everything else."

Jacob Effron

18,374 views • 1 month ago

Liberal man is DEMANDING for ICE agents to be vetted before coming to Minneapolis. “We have no idea who they are, or where they come from..” they literally want to vet federal agents, but not illegal immigrants living in the city. lol

Liberal man is DEMANDING for ICE agents to be vetted before coming to Minneapolis. “We have no idea who they are, or where they come from..” they literally want to vet federal agents, but not illegal immigrants living in the city. lol

American AF 🇺🇸

865,903 views • 4 months ago