正在加载视频...

视频加载失败

AI coding tools are great at generating code. But can an AI agent take a vague idea, understand the whole project, make decisions, write the code, run it, test it, and actually turn it into something usable? I put EvoX Agent through a real-world task to find out. Here’s...

73,008 次观看 • 2 个月前 •via X (Twitter)

38 条评论

Jay Bisen 的头像
Jay Bisen2 个月前

EvoX Agent goes beyond coding; it can research, analyze, execute, and create. It can turn complex tasks into structured workflows and handle them end to end. With reusable skills and multi-agent collaboration, it keeps getting more capable. One agent, multiple workflows, less tool-hopping.

Jay Bisen 的头像
Jay Bisen2 个月前

I gave EvoX Agent a challenge that went beyond simple coding: Build a realistic tornado animation with dynamic movement and effects. It figured out the logic, built the experience, and brought the concept to life without me having to manually connect every piece.

Jay Bisen 的头像
Jay Bisen2 个月前

I wanted to test EvoX on something completely different: Analyze BTC’s price performance throughout July. It processed the data, identified the key price movements, and turned everything into a clear, dynamic chart. Raw market data → actionable insights at a glance.

Jay Bisen 的头像
Jay Bisen2 个月前

Curious what EvoX can do with your own data? Turn raw information into clear insights, charts, and useful outputs. Put it to the test on your next project. Try EvoX Agent Beta ↓

Al Tech World 的头像
Al Tech World2 个月前

AI that can actually execute and validate its own work is a very different proposition from autocomplete.

Sanskriti Naruka 的头像
Sanskriti Naruka2 个月前

Loved seeing EvoX turn a vague idea into a working tornado animation and a clean BTC chart that’s the kind of end-to-end agent that actually ships usable work.

Harsh Pawar 的头像
Harsh Pawar2 个月前

Definitely an interesting direction. The shift from “AI writes code” to “AI completes projects” is happening fast.

Hussain Hashim | Building SundayBack 的头像
Hussain Hashim | Building SundayBack2 个月前

@JayBisen473370 I hit that roadblock building my own thing. Real challenge is balancing AI decisions with user intent. Curious how EvoX Agent handles this!

Minds That Build 的头像
Minds That Build2 个月前

This feels much closer to having an AI teammate than having a simple coding assistant.

The Ant Philosophy 的头像
The Ant Philosophy2 个月前

The BTC analysis example shows how useful agents can be beyond traditional software development.

Daily Wisdom 的头像
Daily Wisdom2 个月前

“Less tool-hopping” is honestly one of the biggest benefits of this approach.

Life Mastery 的头像
Life Mastery2 个月前

The real test for AI isn’t whether it can generate code. It’s whether it can turn an idea into something that actually works.

Shweta singh 的头像
Shweta singh2 个月前

The combination of creativity and data analysis in these examples is pretty compelling.

Wise Philosophy 的头像
Wise Philosophy2 个月前

The ability to research, reason, execute, and iterate in one workflow is where things get interesting.

Amir Ansari 的头像
Amir Ansari2 个月前

Less switching between coding, research, analysis, and visualization could save a lot of time.

Apex Mentality 的头像
Apex Mentality2 个月前

AI agents are becoming less about writing snippets and more about completing actual tasks.

Aiden Tech 的头像
Aiden Tech2 个月前

Great share

marium 的头像
marium2 个月前

That's a great question. Curious to see how EvoX Agent performed on this real-world task!

Mind Yeti 的头像
Mind Yeti2 个月前

I like the idea of testing an agent with completely different types of tasks. That’s a much better benchmark.

Arish Ai 的头像
Arish Ai2 个月前

There’s a big difference between an AI that writes code and one that can actually complete a task.

Ismail Khan 的头像
Ismail Khan2 个月前

Building a sustainable habit routine with rewards and fallbacks sounds like a solid test case. Can't wait to see the final output!

Holland Tech Marco 的头像
Holland Tech Marco2 个月前

That tornado demo is a surprisingly good stress test for an AI agent.

Amaxa AI 的头像
Amaxa AI2 个月前

This is the real test for AI coding agents—not just writing code, but understanding the goal, making decisions, testing, and actually shipping something usable. 🔥 If you want, I can also �⁠make it **shorter, �⁠more viral, or �⁠more natural for X**.

Meenakshi Yadav 的头像
Meenakshi Yadav2 个月前

The BTC chart example is a nice reminder that AI agents can be useful for data tasks too, not just coding.

Manish Kumar Shah 的头像
Manish Kumar Shah2 个月前

The real benchmark will be consistency. Can it produce reliable results across completely different projects?

Onil Coder 的头像
Onil Coder2 个月前

Really impressive. Looking forward to what's next.

Edwin | AI Systems 的头像
Edwin | AI Systems2 个月前

The interesting part isn't the happy path. It's what happens when the agent hits a flaky test or a bad assumption mid-run. Did EvoX recover without a human restart, or was it one clean shot?

Christopher Riley 的头像
Christopher Riley2 个月前

That's great update

0xZubair 🃏 的头像
0xZubair 🃏2 个月前

cool test to see if it can actually ship something

Maria Watson 的头像
Maria Watson2 个月前

We’re moving from “AI can write code” to a much bigger question: “What can AI actually build and complete?”

Aiden Tech 的头像
Aiden Tech2 个月前

Build it, run it, test it” is a very different level of AI capability than simply completing code.

Flix Muller 的头像
Flix Muller2 个月前

Feels like we’re getting closer to AI that can actually own a development workflow.

pushkar soni 的头像
pushkar soni2 个月前

It would be interesting to see the same agent take an existing project, find the problems, and improve it autonomously.

Roxo 的头像
Roxo2 个月前

Noce share

peter 的头像
peter2 个月前

Amazing share

shobhanai 的头像
shobhanai2 个月前

Great share

Mark 的头像
Mark2 个月前

Great post

Dhruv kumar 的头像
Dhruv kumar2 个月前

This is the real test for AI. Most tools can write a snippet, but managing an entire workflow independently is a huge leap. Excited to see how EvoX Agent handled it!

相关视频

Bash is all you need! Which is why I'm introducing my holiday project: just-bash just-bash is a pretty complete implementation of bash in TypeScript designed to be used as a bash tool by AI agents. Because it turns out agents love exploring data via shell scripts, even beyond coding. It comes with grep, sed, awk and the 99th percentile features that an agent like Claude Code or Cursor would use. In fact, Claude Code can use it for secure bash execution. In the package - A bash-tool for AI SDK - A binary for use by yourself or your coding agents - An overlay filesystem to feed files to your agent securely - A Vercel Sandbox compatible API, so you can quickly upgrade to a real VM if you need to run binaries - An example AI agent that explores the just-bash code base using just-bash - I imported the Oils shell bash compatibility suite and just-bash passes a very good chunk What is interesting about this codebase: It was essentially entirely written by Opus 4.5. Coding agents love bash and they are good at reproducing it. They are also great at text-book recursive descent parsers and AST tweet-walk interpreters. That said, it is, like, a lot of code and I didn't read it all 😅. This is very much a hack, but it also seems to be _really_ useful. I haven't really found anything agents want to use that it doesn't support and it's fast and secure (caveats apply). It doesn't have write access to your computer and the filesystem is given a root that the agent cannot escape from. Find it at Related: Our recent blog post how we migrated our data analysis agent to bash tools and achieved incredible quality improvements The video shows the example agent investigating the just-bash code base

Malte Ubl

125,326 次观看 • 10 个月前

Big win for AI Engineers! If you're deploying agents in production, this one is for you. (bookmark it) There's a new system that can test and evaluate your agents against the real world, then write the fixes, all before you deploy them. AI agents look great in demos, until real users start breaking them. Your agent can tell a customer "refund done" on a $1,080 invoice without ever checking who she is, and the logs will still look clean. Normally you only find a failure like that after a real customer runs into it. Plurai Simulation lets your agent fail in a simulated world first, so you can fix the problem before any user sees it. ↳ It rebuilds your actual agent, its tools and its environment from your code ↳ Simulated users push the agent into the edge cases that matter for your use case ↳ Evals built for your agent check the actions it took, like whether it verified the customer before issuing that refund ↳ Every failure comes with the run that exposed it and a fix you can verify ↳ Any production failure can become a simulation you replay against every new version It ships as a single plugin for coding agents like Claude Code and Codex, and it works with whatever agent you're building. I've started running it on my own agents too. It's free for every AI engineer. Link in the first comment. I worked with the Plurai team to put this together. Here's the whole system, step by step, from the first simulated user to a re-tested fix:

Akshay 🚀

10,728 次观看 • 5 天前