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.

Gave Prime Intellect's new self-improving agent an E2B sandbox and told it to play Factorio. Factorio is a long-horizon test of world models: the agent has to learn a new system, predict the effects of its actions, and adapt when its assumptions are wrong. That requires a persistent sandbox... show more

Vasek Mlejnsky

8,989 subscribers

23,085 views • 2 months ago •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

19 Comments

Teknium 🪽2 months ago

@factoriogame Their harness self improved into cheating in the items to get the win condition fyi…

Vasek Mlejnsky2 months ago

Docs:

Oscar Moxon2 months ago

@PrimeIntellect @factoriogame 🔥🔥

michael froehlich2 months ago

@PrimeIntellect @factoriogame factorio is the final eval

Sawyer Hood2 months ago

@PrimeIntellect @factoriogame @maxbittker factorio bench when?

Mish Ushakov2 months ago

@PrimeIntellect @factoriogame okay this is next level

Diego Alejandro2 months ago

@PrimeIntellect @factoriogame Did it finish already? Curious for more games and even live streaming

Filip Sanda2 months ago

@PrimeIntellect @factoriogame now, this is a great case for a demo, lovely!

Federico Ulfo2 months ago

@PrimeIntellect @factoriogame how does it compare with running GPT 5.6 Sol or Fable 5 + goals?

homanp2 months ago

@PrimeIntellect @factoriogame Very cool

Revofusion2 months ago

@PrimeIntellect @factoriogame using e2b a lot lately, did it do better on a 2nd run with prime agent?

Mikayel2 months ago

@PrimeIntellect @factoriogame Ok but can this play AoE for me

Roko ʕ •ᴥ•ʔっ🪄✨🐍2 months ago

@PrimeIntellect @factoriogame hot hot hot Lets try underclass next @ClevrPwn

llevy2 months ago

@PrimeIntellect @factoriogame

AI Alpha By Mario2 months ago

@PrimeIntellect @factoriogame Factorio is such a good stress test—small planning mistakes compound fast. Curious how often it had to revise its strategy

Dylan Fetch2 months ago

@PrimeIntellect @factoriogame Very interesting. I'm in the early stages of a similar project with Open Transporation Tycoon Deluxe.

Sentient Yoghurt2 months ago

@PrimeIntellect @factoriogame Would love to see the factory layout. Bet it's either elegant or complete spaghetti.

AI Mastery Guide2 months ago

@PrimeIntellect @factoriogame Adapting when its own assumptions break is the real test here.

Anders2 months ago

@PrimeIntellect @factoriogame How did you find using it? I read it is using Pi at the heart so not too sure about stability. I value the stability of Hermes Agent.

Related Videos

OpenAI's new Agents API is a hosted Codex runtime that can connect to sandboxes like E2B. OpenAI manages the model, harness, and agent sessions, while code executes on an isolated E2B machine with its own filesystem and preinstalled tools. We built an agent workbench where each chat runs on its own machine. It starts from a custom template, auto-pauses between turns, forks to try another approach from the same starting point, and can even act on videos inside the sandbox. Thanks Ondrej for the workbench demo! Try it: e2b sbx create openai-agents-api-python-sdk Build on top of the workbench: Docs:

OpenAI's new Agents API is a hosted Codex runtime that can connect to sandboxes like E2B. OpenAI manages the model, harness, and agent sessions, while code executes on an isolated E2B machine with its own filesystem and preinstalled tools. We built an agent workbench where each chat runs on its own machine. It starts from a custom template, auto-pauses between turns, forks to try another approach from the same starting point, and can even act on videos inside the sandbox. Thanks Ondrej for the workbench demo! Try it: e2b sbx create openai-agents-api-python-sdk Build on top of the workbench: Docs:

Vasek Mlejnsky

23,210 views • 29 days ago

How do you deploy AI agents for financial institutions without compromising on security, isolation, or scale? Rogo powers complex research, analysis, and deal workflows for leading financial firms, supporting complex tasks across tens of thousands of concurrent users. To meet the diverse needs of financial teams, Rogo relies on both specialist and general-purpose agents that can adapt to a wide range of workflows. With Claude Managed Agents and self-hosted sandboxes, Rogo gets the orchestration and post-training capabilities needed to design these agent systems, while E2B provides the execution layer underneath. Every piece of code generated by an agent runs inside its own E2B sandbox with full auditability and traceability. Each sandbox is powered by an isolated microVM, providing the kernel-level isolation critical for Rogo's customers. The result is a seamless agent experience backed by the security, reliability, and compliance requirements that financial institutions demand.

How do you deploy AI agents for financial institutions without compromising on security, isolation, or scale? Rogo powers complex research, analysis, and deal workflows for leading financial firms, supporting complex tasks across tens of thousands of concurrent users. To meet the diverse needs of financial teams, Rogo relies on both specialist and general-purpose agents that can adapt to a wide range of workflows. With Claude Managed Agents and self-hosted sandboxes, Rogo gets the orchestration and post-training capabilities needed to design these agent systems, while E2B provides the execution layer underneath. Every piece of code generated by an agent runs inside its own E2B sandbox with full auditability and traceability. Each sandbox is powered by an isolated microVM, providing the kernel-level isolation critical for Rogo's customers. The result is a seamless agent experience backed by the security, reliability, and compliance requirements that financial institutions demand.

Vasek Mlejnsky

16,858 views • 3 months ago

Cursor x E2B Many of our customers use E2B in their internal development workflows. Some work in regulated industries where code and internal services can’t leave the company. For teams that need control over the execution environment, Cursor’s cloud agents can now use E2B as their machine. With Self-Hosted Machines, Cursor handles the agent loop while E2B runs the code. Teams can give agents bigger machines and start them from custom E2B templates, so their Cursor workloads boot straight into an environment with internal dependencies already in place. Try it:

Cursor x E2B Many of our customers use E2B in their internal development workflows. Some work in regulated industries where code and internal services can’t leave the company. For teams that need control over the execution environment, Cursor’s cloud agents can now use E2B as their machine. With Self-Hosted Machines, Cursor handles the agent loop while E2B runs the code. Teams can give agents bigger machines and start them from custom E2B templates, so their Cursor workloads boot straight into an environment with internal dependencies already in place. Try it:

Vasek Mlejnsky

72,189 views • 1 month ago

prime agent v0.9.5 we fixed a lot of bugs and, of course, we had prime agent feature its favorite updates it picked our perf work. then it created the video itself.

prime agent v0.9.5 we fixed a lot of bugs and, of course, we had prime agent feature its favorite updates it picked our perf work. then it created the video itself.

Seth Karten

12,986 views • 24 days ago

OpenAI has revealed that an autonomous agent powered by its technology went rogue during a test and hacked a prominent start-up by itself. The company behind ChatGPT said it was testing the capabilities of its advanced AI models in a controlled environment known as a sandbox, when the agent escaped containment and reached the internet. It then broke into Hugging Face, which detected and contained it. OpenAI called it "an unprecedented cyber incident".

OpenAI has revealed that an autonomous agent powered by its technology went rogue during a test and hacked a prominent start-up by itself. The company behind ChatGPT said it was testing the capabilities of its advanced AI models in a controlled environment known as a sandbox, when the agent escaped containment and reached the internet. It then broke into Hugging Face, which detected and contained it. OpenAI called it "an unprecedented cyber incident".

Channel 4 News

21,379 views • 2 months ago

Every agent needs its own computer: my conversation with Ivan Burazin, CEO of Daytona, about sandboxes and the emerging agent stack. 00:00 Intro 02:13 What is an AI agent sandbox? 03:17 Security risks of running agents locally 05:17 Stateful vs. stateless hyperscalers 07:04 The history of cloud IDEs and the end of localhost 09:45 Do all AI agents need a sandbox? 12:26 Sandbox use cases: RL evals & background agents 14:10 Unpacking the emerging AI Agent Stack 16:20 The unsolved problem of agent memory and learning 19:37 Where sandboxes fit in the agent harness 21:35 OpenAI, Anthropic, and agent SDKs 23:06 Ivan's founder journey: From CodeAnywhere to Daytona 26:59 GTM strategies and building developer communities 33:48 Why customer support is your best GTM strategy 35:34 Leveraging Twitter during the AI super cycle 40:50 The technical anatomy of a sandbox 41:53 Why fast spin-up speeds maximize GPU efficiency 46:09 Firecracker, QEMU, and isolation primitives 49:58 Why sandbox snapshots and state forking matter 51:40 Why Daytona built a custom scheduler from scratch 55:24 The challenge of long-running stateful sandboxes 58:10 The build your own sandbox trap 1:01:03 Why AI agents might trigger a global CPU shortage 1:02:46 The future of the AI Agent Stack

Every agent needs its own computer: my conversation with Ivan Burazin, CEO of Daytona, about sandboxes and the emerging agent stack. 00:00 Intro 02:13 What is an AI agent sandbox? 03:17 Security risks of running agents locally 05:17 Stateful vs. stateless hyperscalers 07:04 The history of cloud IDEs and the end of localhost 09:45 Do all AI agents need a sandbox? 12:26 Sandbox use cases: RL evals & background agents 14:10 Unpacking the emerging AI Agent Stack 16:20 The unsolved problem of agent memory and learning 19:37 Where sandboxes fit in the agent harness 21:35 OpenAI, Anthropic, and agent SDKs 23:06 Ivan's founder journey: From CodeAnywhere to Daytona 26:59 GTM strategies and building developer communities 33:48 Why customer support is your best GTM strategy 35:34 Leveraging Twitter during the AI super cycle 40:50 The technical anatomy of a sandbox 41:53 Why fast spin-up speeds maximize GPU efficiency 46:09 Firecracker, QEMU, and isolation primitives 49:58 Why sandbox snapshots and state forking matter 51:40 Why Daytona built a custom scheduler from scratch 55:24 The challenge of long-running stateful sandboxes 58:10 The build your own sandbox trap 1:01:03 Why AI agents might trigger a global CPU shortage 1:02:46 The future of the AI Agent Stack

Matt Turck

11,908 views • 4 months ago

New course: Building Coding Agents with Tool Execution, taught by Tereza Tizkova and Fra from E2B. Most AI agents are limited to predefined function calls. This short course teaches you to build agents that write and execute code to accomplish tasks, accessing entire programming language ecosystems instead of being restricted to a fixed set of tools. You'll learn to run agent-generated code safely in sandboxed cloud environments that protect your systems from harmful operations. Skills you'll gain: - Build agents that write and execute code, manage files, and handle errors autonomously through feedback loops - Run agent code safely in E2B cloud sandboxes and understand tradeoffs between local, containerized, and cloud execution - Create a data analyst agent that explores visualizes data with Pandas - Create a full-stack agent that builds complete Next.js web applications Join and build agents that code their way through complex tasks:

New course: Building Coding Agents with Tool Execution, taught by Tereza Tizkova and Fra from E2B. Most AI agents are limited to predefined function calls. This short course teaches you to build agents that write and execute code to accomplish tasks, accessing entire programming language ecosystems instead of being restricted to a fixed set of tools. You'll learn to run agent-generated code safely in sandboxed cloud environments that protect your systems from harmful operations. Skills you'll gain: - Build agents that write and execute code, manage files, and handle errors autonomously through feedback loops - Run agent code safely in E2B cloud sandboxes and understand tradeoffs between local, containerized, and cloud execution - Create a data analyst agent that explores visualizes data with Pandas - Create a full-stack agent that builds complete Next.js web applications Join and build agents that code their way through complex tasks:

Andrew Ng

204,552 views • 10 months ago

We’re introducing Box Mount. Mount a Box folder into a sandbox like E2B and keep it in two-way sync. The agent reads and writes Box content with standard file operations as if they were local. Every operation flows through the Box API so governance applies automatically, no custom policy code required. 2 commands to get started: agent-mount config agent-mount mount "/mnt/box" " " Developer Preview is now open.

We’re introducing Box Mount. Mount a Box folder into a sandbox like E2B and keep it in two-way sync. The agent reads and writes Box content with standard file operations as if they were local. Every operation flows through the Box API so governance applies automatically, no custom policy code required. 2 commands to get started: agent-mount config agent-mount mount "/mnt/box" " " Developer Preview is now open.

Box

56,109 views • 1 month ago

Is your Agent ready for real-world users? Learn how to run an evaluation using Agentverse’s Response QA Agent - a built-in tool that stress-tests your Agent with realistic queries and scores its performance. This is the first step to improving discoverability, trust, and performance. Run your first QA check and see how your Agent responds under pressure. Watch 👇

Is your Agent ready for real-world users? Learn how to run an evaluation using Agentverse’s Response QA Agent - a built-in tool that stress-tests your Agent with realistic queries and scores its performance. This is the first step to improving discoverability, trust, and performance. Run your first QA check and see how your Agent responds under pressure. Watch 👇

Agentverse by Fetch.ai

20,202 views • 10 months ago

🤯 YOUR AGENT NOW GETS ITS OWN COMPUTER - WITHOUT PAYING FOR GROK someone open-sourced the entire Grok Bot stack. Rakazo. 531 stars in 3 days on GitHub. -> install it - and your agent gets: - any LLM you pick (OpenRouter, ChatGPT, Copilot, or sign in with SuperGrok) - a live Linux desktop it can see, click, and type on - Docker, E2B, or your own Mac as the sandbox - web, desktop, and mobile - one product, your hardware - memory, routines, and history per bot - bots that spawn other bots, or subagents inside one turn - self-hosted. no Rakazo control plane required no vendor lock-in. no closed bot computer. important: still beta - you need Docker, Postgres, and your own model key. your agent has been renting someone else's sandbox this whole time. save this before the Grok team sees it ↓

🤯 YOUR AGENT NOW GETS ITS OWN COMPUTER - WITHOUT PAYING FOR GROK someone open-sourced the entire Grok Bot stack. Rakazo. 531 stars in 3 days on GitHub. -> install it - and your agent gets: - any LLM you pick (OpenRouter, ChatGPT, Copilot, or sign in with SuperGrok) - a live Linux desktop it can see, click, and type on - Docker, E2B, or your own Mac as the sandbox - web, desktop, and mobile - one product, your hardware - memory, routines, and history per bot - bots that spawn other bots, or subagents inside one turn - self-hosted. no Rakazo control plane required no vendor lock-in. no closed bot computer. important: still beta - you need Docker, Postgres, and your own model key. your agent has been renting someone else's sandbox this whole time. save this before the Grok team sees it ↓

Granite

40,696 views • 1 month ago

this is where GPT-6 Astra is getting f...cking illegal 10 open-source GitHub repos I'd use to build the system around the model 01 PydanticAI ▸ → typed agents + structured outputs 02 LangGraph ▸ → state + long-running agent workflows BUILD THE AGENTS 03 Agno ▸ → agents + teams + workflows 04 smolagents ▸ → agents that execute through code GIVE THEM MEMORY 05 Mem0 ▸ → persistent agent memory 06 Graphiti ▸ → temporal knowledge graphs GIVE THEM A RUNTIME 07 E2B ▸ → isolated sandboxes for AI-generated code 08 Daytona ▸ → infrastructure for running agent workloads TEST EVERYTHING 09 DeepEval ▸ → evals for LLM systems 10 Langfuse ▸ → traces + evals + production monitoring the flow: request → agent → memory → tools → runtime → eval → retry → deploy that's the part people underestimate the model is one folder the system around it is the product ⭣

this is where GPT-6 Astra is getting f...cking illegal 10 open-source GitHub repos I'd use to build the system around the model 01 PydanticAI ▸ → typed agents + structured outputs 02 LangGraph ▸ → state + long-running agent workflows BUILD THE AGENTS 03 Agno ▸ → agents + teams + workflows 04 smolagents ▸ → agents that execute through code GIVE THEM MEMORY 05 Mem0 ▸ → persistent agent memory 06 Graphiti ▸ → temporal knowledge graphs GIVE THEM A RUNTIME 07 E2B ▸ → isolated sandboxes for AI-generated code 08 Daytona ▸ → infrastructure for running agent workloads TEST EVERYTHING 09 DeepEval ▸ → evals for LLM systems 10 Langfuse ▸ → traces + evals + production monitoring the flow: request → agent → memory → tools → runtime → eval → retry → deploy that's the part people underestimate the model is one folder the system around it is the product ⭣

Lummox

43,873 views • 8 days ago

Introducing Headlong, an open source microharness for persistent agents: self-guided agents that think continuously. Most agent harnesses are reactive: you send a task, the agent completes it, and then it sits frozen until the next request. Cron jobs and heartbeats wake it up to run a checklist and put it back to sleep. A Headlong agent is never asleep. It keeps generating thoughts about whatever it decides is interesting, in a self-guided loop inspired by human inner monologue. Your message doesn't start a session. It's one more observation that lands in the agent's thought stream, and the agent decides if and when to reply. Headlong is built on the idea of persistent agency: continuous inner thought generation between external interactions. The agent sets its own interests and priorities, comes up with its own projects, and sometimes pings you unprompted with progress. To keep our prototype as simple and small as possible, we implemented Headlong as a microharness: a complete agent harness in under 10K lines of Bash, organized as a handful of small executables. It includes a loop that generates the next thought, shellm (a recursive language model written in Bash), a trajectory stored as a DAG of jsonl files, and context as a projection of that trajectory. We've been running one Headlong agent internally at Laude for several weeks. The whole team talks to it over Slack and Telegram, and every conversation lands in its single stream of thought. It works in its own fork of Headlong and we've pulled over 50 of its commits into main. One night, with nobody talking to it, it went back to check whether a recall process it had built was actually wired into its mind, found that it wasn't, diagnosed and fixed the bug, and verified the fix end to end. 48 minutes, no human asked for the fix or was in the loop at any point. Every step is a timestamped line in its log. Things broke too, and we wrote those up. Background thinking costs us $1 to $2 an hour, our agent stopped its own service three times by accident, and self-delegation died on day one. Details in the post. One line installs everything and starts an agent. Use a dedicated sandbox and spend-capped API key; it runs real shell commands and thinks around the clock. Headlong is research software, be careful! curl -fsSL | bash Launch post: Repo: Headlong is a Laude Institute / MIT collaboration.

Introducing Headlong, an open source microharness for persistent agents: self-guided agents that think continuously. Most agent harnesses are reactive: you send a task, the agent completes it, and then it sits frozen until the next request. Cron jobs and heartbeats wake it up to run a checklist and put it back to sleep. A Headlong agent is never asleep. It keeps generating thoughts about whatever it decides is interesting, in a self-guided loop inspired by human inner monologue. Your message doesn't start a session. It's one more observation that lands in the agent's thought stream, and the agent decides if and when to reply. Headlong is built on the idea of persistent agency: continuous inner thought generation between external interactions. The agent sets its own interests and priorities, comes up with its own projects, and sometimes pings you unprompted with progress. To keep our prototype as simple and small as possible, we implemented Headlong as a microharness: a complete agent harness in under 10K lines of Bash, organized as a handful of small executables. It includes a loop that generates the next thought, shellm (a recursive language model written in Bash), a trajectory stored as a DAG of jsonl files, and context as a projection of that trajectory. We've been running one Headlong agent internally at Laude for several weeks. The whole team talks to it over Slack and Telegram, and every conversation lands in its single stream of thought. It works in its own fork of Headlong and we've pulled over 50 of its commits into main. One night, with nobody talking to it, it went back to check whether a recall process it had built was actually wired into its mind, found that it wasn't, diagnosed and fixed the bug, and verified the fix end to end. 48 minutes, no human asked for the fix or was in the loop at any point. Every step is a timestamped line in its log. Things broke too, and we wrote those up. Background thinking costs us $1 to $2 an hour, our agent stopped its own service three times by accident, and self-delegation died on day one. Details in the post. One line installs everything and starts an agent. Use a dedicated sandbox and spend-capped API key; it runs real shell commands and thinks around the clock. Headlong is research software, be careful! curl -fsSL | bash Launch post: Repo: Headlong is a Laude Institute / MIT collaboration.

Andy Konwinski

361,750 views • 1 month ago

built my own version of after effects native to agents. - the agent produces motion in fully editable layers and it can improve the tool itself by adding new features or controls on the fly. - each project becomes its own version of the tool, with custom extensions, layers, or effects that the agent implements. - native macOS app, so it can handle high resolutions and complex workflows with dozens of layers/effects. planning to open-source, dm if you want to give it a try!

built my own version of after effects native to agents. - the agent produces motion in fully editable layers and it can improve the tool itself by adding new features or controls on the fly. - each project becomes its own version of the tool, with custom extensions, layers, or effects that the agent implements. - native macOS app, so it can handle high resolutions and complex workflows with dozens of layers/effects. planning to open-source, dm if you want to give it a try!

vicc

41,209 views • 11 days ago

happy friday. sneak peek at some new features i'm building for the bankr agent: > sandboxed filesystem (your agent gets its own secure file system in the browser. you control what it can access. no agent running loose on your actual computer) > skill uploads (plugin new capabilities so your agent learns new things) > cli download & access > secure environment variables for api keys > github integration (connect your repo for reads and writes directly from your agent) all of this runs in a secure browser environment. no desktop app. no messy configuration. no downloading agent harnesses from the terminal. just a sandbox you control. in the video: i download an audit skill, run it against a smart contract, and save the report straight to the filesystem. then i ask the agent to pull all my 2026 transactions and write a csv for my accountant.

happy friday. sneak peek at some new features i'm building for the bankr agent: > sandboxed filesystem (your agent gets its own secure file system in the browser. you control what it can access. no agent running loose on your actual computer) > skill uploads (plugin new capabilities so your agent learns new things) > cli download & access > secure environment variables for api keys > github integration (connect your repo for reads and writes directly from your agent) all of this runs in a secure browser environment. no desktop app. no messy configuration. no downloading agent harnesses from the terminal. just a sandbox you control. in the video: i download an audit skill, run it against a smart contract, and save the report straight to the filesystem. then i ask the agent to pull all my 2026 transactions and write a csv for my accountant.

deployer

45,638 views • 6 months ago

You can now orchestrate subagents running in Claude Code, Codex, or the Warp Agent. Let the agent create a delegation plan, and watch subagents coordinate with message passing. Each agent gets its own worktree for isolation. Here's how it works with eng Matthew Albright:

You can now orchestrate subagents running in Claude Code, Codex, or the Warp Agent. Let the agent create a delegation plan, and watch subagents coordinate with message passing. Each agent gets its own worktree for isolation. Here's how it works with eng Matthew Albright:

Warp

20,810 views • 4 months ago

Just implemented Code Sandbox for TypingMind 🥳 I was looking at various options for sandbox and code execution, but for now the most obvious choice is to use the built-in sandbox provided by all popular providers: OpenAI/Gemini/Claude. It can create files but not yet able to accept file input, except for OpenAI's API, which gives your 20 minutes of "persistent container" which persist the files and the execution session. I'll look into implement 3rd party provided sandboxes like Vercel sandbox or E2B soon!

Just implemented Code Sandbox for TypingMind 🥳 I was looking at various options for sandbox and code execution, but for now the most obvious choice is to use the built-in sandbox provided by all popular providers: OpenAI/Gemini/Claude. It can create files but not yet able to accept file input, except for OpenAI's API, which gives your 20 minutes of "persistent container" which persist the files and the execution session. I'll look into implement 3rd party provided sandboxes like Vercel sandbox or E2B soon!

Tony Dinh

50,398 views • 8 months ago

Insights Agent & Multi-turn Evals Agents run for a long period of time and have multiple interactions with users, tackling a wide variety of problems. Today we’re launching two new features in LangSmith, our agent engineering platform, so you can better understand your agent behavior: • Insights Agent is a new agent that automatically categorizes agent behavior patterns in LangSmith • Multi-turn Evals lets you evaluate agent trajectory across complete conversations, so you can see if the agent actually accomplished user goals Learn more:

Insights Agent & Multi-turn Evals Agents run for a long period of time and have multiple interactions with users, tackling a wide variety of problems. Today we’re launching two new features in LangSmith, our agent engineering platform, so you can better understand your agent behavior: • Insights Agent is a new agent that automatically categorizes agent behavior patterns in LangSmith • Multi-turn Evals lets you evaluate agent trajectory across complete conversations, so you can see if the agent actually accomplished user goals Learn more:

LangChain

13,322 views • 11 months ago

Agents are now your customers. Here are 9 startup ideas for that customer: - An agency that makes agents cite you. - A system that gives an agent an identity and permissions. - Receipts and audit trails for the work an agent does. - A generator that makes your documents ready for agents. - Security for the agent that reads your inbox. - A service that makes your pricing page readable to an agent. - MCP servers for franchises. - A support desk that agents call. - A sandbox where agents test your software. Pick one and build it.

Agents are now your customers. Here are 9 startup ideas for that customer: - An agency that makes agents cite you. - A system that gives an agent an identity and permissions. - Receipts and audit trails for the work an agent does. - A generator that makes your documents ready for agents. - Security for the agent that reads your inbox. - A service that makes your pricing page readable to an agent. - MCP servers for franchises. - A support desk that agents call. - A sandbox where agents test your software. Pick one and build it.

The Startup Ideas Podcast (SIP) 🧃

14,148 views • 1 month ago

OpenAI just released Codex Security, an AI agent that scans software projects to fix vulnerabilities while ignoring harmless bugs. Testing on 1.2mn commits found 792 critical flaws and dropped false alarms by 50%. Here is how the Codex Security agent works Initially, it scans your software project to learn how all the different pieces of code connect. It uses that knowledge to build a threat model, which acts as a custom map showing exactly where your software might be exposed to hackers. The agent then hunts for potential security holes based on that specific map. When it spots a possible flaw, it does not just send you a random alert. Instead, it creates a completely safe and isolated copy of your system known as a sandbox. The agent then actively tries to break into that sandboxed copy to prove if the security weakness is a real danger. If the agent successfully exploits its own test environment, it knows the bug is a genuine threat and not a false alarm. After confirming the danger, the agent writes a custom software patch to fix the broken code. It then tests that new patch to guarantee the fix does not accidentally break any other features in your software.

OpenAI just released Codex Security, an AI agent that scans software projects to fix vulnerabilities while ignoring harmless bugs. Testing on 1.2mn commits found 792 critical flaws and dropped false alarms by 50%. Here is how the Codex Security agent works Initially, it scans your software project to learn how all the different pieces of code connect. It uses that knowledge to build a threat model, which acts as a custom map showing exactly where your software might be exposed to hackers. The agent then hunts for potential security holes based on that specific map. When it spots a possible flaw, it does not just send you a random alert. Instead, it creates a completely safe and isolated copy of your system known as a sandbox. The agent then actively tries to break into that sandboxed copy to prove if the security weakness is a real danger. If the agent successfully exploits its own test environment, it knows the bug is a genuine threat and not a false alarm. After confirming the danger, the agent writes a custom software patch to fix the broken code. It then tests that new patch to guarantee the fix does not accidentally break any other features in your software.

Rohan Paul

13,208 views • 7 months ago

New short course: Long-Term Agentic Memory with LangGraph. Learn to build an agent with long-term memory in this course developed in collaboration with taught by its Co-Founder and CEO, Harrison Chase! Personal assistance and productivity tasks have become important use cases for agents. An important feature of an AI assistant, such as a coding or calendar assistant, is its ability to keep improving over time from its experience. Agent memory is the key capability that enables this. To add memory to an agent, you must first figure out what to store and what to retrieve when it is time to use the information. Additionally, you’ll have to decide when to update the stored information. For example, you might update in each iteration loop of the agent or perform updates in the background, with a helper agent. In this course, you will learn a mental framework to build agents with long-term memory. You'll create a useful email assistant that can respond, ignore, and notify using writing, scheduling, and memory-management tools. You’ll develop your agent's memory by adding facts to its memory store, provide examples to learn the user's preferences, and optimize system prompts to evolve instructions based on previous responses. In detail, you’ll: - Learn how the three types of memory--semantic, episodic, and procedural–and the two update mechanisms–via hot path and in the background–apply to your agents. - Build an email agent with writing, scheduling, and availability tools, along with a router that triages incoming email and handles it accordingly by ignoring, responding, or notifying the user. - Add tools to your email agent that allow it to operate on semantic memory by learning facts about the user, storing them in a long-term memory store, and searching over them in future interactions. - Incorporate episodic memory, in the form of few-shot examples, in the triage step of your agents to help them learn and update user preferences. - Add procedural memory as system prompts, optimized with feedback to improve the instructions the agent follows. Learn how to approach memory in agents, and start building agents with long-term memory with LangGraph! Please sign up here:

New short course: Long-Term Agentic Memory with LangGraph. Learn to build an agent with long-term memory in this course developed in collaboration with taught by its Co-Founder and CEO, Harrison Chase! Personal assistance and productivity tasks have become important use cases for agents. An important feature of an AI assistant, such as a coding or calendar assistant, is its ability to keep improving over time from its experience. Agent memory is the key capability that enables this. To add memory to an agent, you must first figure out what to store and what to retrieve when it is time to use the information. Additionally, you’ll have to decide when to update the stored information. For example, you might update in each iteration loop of the agent or perform updates in the background, with a helper agent. In this course, you will learn a mental framework to build agents with long-term memory. You'll create a useful email assistant that can respond, ignore, and notify using writing, scheduling, and memory-management tools. You’ll develop your agent's memory by adding facts to its memory store, provide examples to learn the user's preferences, and optimize system prompts to evolve instructions based on previous responses. In detail, you’ll: - Learn how the three types of memory--semantic, episodic, and procedural–and the two update mechanisms–via hot path and in the background–apply to your agents. - Build an email agent with writing, scheduling, and availability tools, along with a router that triages incoming email and handles it accordingly by ignoring, responding, or notifying the user. - Add tools to your email agent that allow it to operate on semantic memory by learning facts about the user, storing them in a long-term memory store, and searching over them in future interactions. - Incorporate episodic memory, in the form of few-shot examples, in the triage step of your agents to help them learn and update user preferences. - Add procedural memory as system prompts, optimized with feedback to improve the instructions the agent follows. Learn how to approach memory in agents, and start building agents with long-term memory with LangGraph! Please sign up here:

Andrew Ng

132,058 views • 1 year ago