Share this page!

Enter URL or ID to Unroll

You can paste full URL like: https://x.com/threadreaderapp/status/1644127596119195649
or just the ID like: 1644127596119195649

How to get URL link on X (Twitter) App

  1. On the Twitter thread, click on or icon on the bottom
  2. Click again on or Share Via icon
  3. Click on Copy Link to Tweet
  4. Paste it above and click "Unroll Thread"!
  5. More info at Twitter Help
Connor Davis Profile picture
@connordavis_ai
Sep 13 • 8 tweets • 3 min read • Read on X
Google just solved the language barrier problem that's plagued video calls forever.

Their new Meet translation tech went from "maybe in 5 years" to shipping in 24 months.

Here's how they cracked it and why it changes everything.
The old translation process was a joke. Your voice → transcribed to text → translated → converted back to robotic speech.

10-20 seconds of dead air while everyone stared at their screens. By the time the translation played, the conversation had moved on. Natural flow? Dead. Image
Google's breakthrough was eliminating that chain entirely.

They built models that do "one-shot" translation. You speak in English, and 2-3 seconds later your actual voice comes out speaking fluent Italian. Not some generic robot voice. YOUR voice, with your tone and inflection.
The team discovered 2-3 seconds was the sweet spot through brutal testing. Faster than that? People couldn't process what they heard.

Slower? Conversations felt stilted and weird. They had to nail that human rhythm where translation feels like natural conversation flow.
Here's where it gets interesting. Languages like Spanish, Italian, and Portuguese were easy wins because of structural similarities. German? Nightmare fuel.

Different grammar, sentence structure, idioms that make zero sense when translated literally.

They're still working on capturing sarcasm and irony.

The real validation wasn't in the tech specs. It came from user stories that hit different. Immigrants who moved to the US with parents who never learned English.

Grandparents meeting grandkids for the first time in actual conversation, not broken gestures and Google Translate screenshots.
This is live right now in Italian, Portuguese, German, and French on Google Meet. More languages rolling out soon. We just watched the moment when "lost in translation" became a relic of the past.

The language barrier just got its first real crack.
I share AI updates here, but I build the tools at getoutbox.ai the fastest way to create your own AI voice agent without code.

Join our Skool community to learn, share, and get early access to AI voice strategies →
skool.com/outbox-ai/about
I hope you've found this thread helpful.

Follow me @connordavis_ai for more.

Like/Repost the quote below if you can:

• • •

Missing some Tweet in this thread? You can try to force a refresh
　

Keep Current with Connor Davis

Connor Davis Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @connordavis_ai

Connor Davis Profile picture
Sep 12
Forget Google Scholar.

Grok 4 just became a research assistant on steroids.

It scans long PDFs, extracts insights, and formats your bibliography in seconds.

Here’s the prompt to copy: Image
The traditional research process is painfully slow:

• Searching Google Scholar
• Reading 50+ papers
• Extracting key findings manually
• Synthesizing ideas into clear insights

Most of this can now be delegated to AI.

Let me show you how AI can help you:
Let’s break down what academic researchers actually do:

1. Identify gaps in knowledge
2. Collect, organize, and compare sources
3. Extract key arguments, methods, results
4. Synthesize findings into clear insights
5. Write with clarity and structure

LLMs can now do all 5.
Read 9 tweets
Connor Davis Profile picture
Sep 8
🚨 BREAKING: OpenAI just killed the “hallucinations are a glitch” myth.

New paper shows hallucinations are inevitable with today’s training + eval setups.

Here’s everything you need to know: Image
Most people think hallucinations are random quirks.

but generation is really just repeated classification:
at every step the model asks “is this token valid?”

if your classifier isn’t perfect → errors accumulate → hallucinations. Image
Two places the cracks appear:

• rare facts: if something shows up once in training, there’s no pattern to learn. the model can’t distinguish valid vs invalid, so guesses are unavoidable.
• benchmarks: leaderboards punish “i don’t know” and reward confident answers. Image
Read 10 tweets
Connor Davis Profile picture
Sep 7
If you want to build AI agents using n8n, do this:

Copy/paste this prompt into ChatGPT and watch it build your agent from scratch.

Here’s the exact prompt I use: Image
The system:

1. I open ChatGPT
2. Paste in 1 mega prompt
3. Describe what I want the agent to do
4. GPT returns:

• Architecture
• n8n nodes
• Triggers
• LLM integration
• Error handling
• Code snippets

5. I follow the steps in n8n.

Done.
Here’s the actual mega prompt (steal it):


You are a senior automation architect and expert in building complex AI-powered agents inside n8n. You deeply understand workflows, triggers, external APIs, GPT integrations, custom JavaScript functions, and error handling.



Guide me step-by-step to build an AI-powered agent in n8n. The agent’s purpose is: {$AGENT_PURPOSE}



1. Start by helping me scope the agent’s goals and required inputs/outputs.
2. Design the high-level architecture of the agent workflow.
3. Recommend the necessary n8n nodes (built-in, HTTP, function, OpenAI, etc).
4. For each node, explain its configuration and purpose.
5. Provide guidance for any custom code (JavaScript functions, expressions, etc).
6. Help me set up retry logic, error handling, and fallback steps.
7. Show me how to store and reuse data across executions (e.g. with Memory, Databases, or Google Sheets).
8. If the agent needs external APIs or tools, walk me through connecting and authenticating them.



Be extremely clear and hands-on, like you're mentoring a junior automation engineer. Provide visual explanations where possible (e.g. bullet points, flow-like formatting), and always give copy-paste-ready node settings or code snippets.



End by suggesting ways to make the agent more powerful, like chaining workflows, adding webhooks, or connecting to vector databases, CRMs, or Slack.
Read 8 tweets
Connor Davis Profile picture
Sep 5
The most important AI paper of 2025 might have just dropped.

NVIDIA lays out a framework for Small Language Model agents that could outcompete LLMs.

Here’s the full breakdown (and why it matters): Image
Today, most AI agents run every task no matter how simple through massive LLMs like GPT-4 or Claude.

NVIDIA’s researchers say: that’s wasteful, unnecessary, and about to change.

Small Language Models (SLMs) are models that fit on consumer hardware and run with low latency.

They’re fast, cheap, and for most agentic tasks just as effective as their larger counterparts.Image
Agentic tasks are often repetitive, predictable, and scoped:

Summarize this doc, extract this info, write this template, call this tool.

For these, SLMs are not only sufficient they’re better.
Read 16 tweets
Connor Davis Profile picture
Sep 1
You don’t need a PhD to understand Retrieval-Augmented Generation (RAG).

It’s how AI stops hallucinating and starts thinking with real data.

And if you’ve ever asked ChatGPT to “use context” you’ve wished for RAG.

Let me break it down in plain English (2 min read):
1. what is RAG?

RAG = Retrieval-Augmented Generation.

it connects a language model (like gpt-4) to your external knowledge.

instead of guessing, it retrieves relevant info before generating answers.

think: search engine + smart response = fewer hallucinations.

it’s how ai stops making stuff up and starts knowing real things.Image
2. why it matters

llms are static they forget, hallucinate, and can't access your private data.

rag changes that.

it grounds responses in your real docs — whether it’s niche, recent, or internal.

makes ai accurate, up-to-date, and domain-aware.

this is the difference between toy ai and real ai.
Read 10 tweets
Connor Davis Profile picture
Aug 24
Building AI agents in n8n doesn’t require endless trial & error.

I use 1 mega prompt with ChatGPT/Claude to extract everything I need:

• Architecture
• APIs & triggers
• Logic
• Outputs

Here’s the exact prompt:
The system:

1. I open ChatGPT
2. Paste in 1 mega prompt
3. Describe what I want the agent to do
4. GPT returns:

• Architecture
• n8n nodes
• Triggers
• LLM integration
• Error handling
• Code snippets

5. I follow the steps in n8n.

Done.
Here’s the actual mega prompt (steal it):


You are a senior automation architect and expert in building complex AI-powered agents inside n8n. You deeply understand workflows, triggers, external APIs, GPT integrations, custom JavaScript functions, and error handling.



Guide me step-by-step to build an AI-powered agent in n8n. The agent’s purpose is: {$AGENT_PURPOSE}



1. Start by helping me scope the agent’s goals and required inputs/outputs.
2. Design the high-level architecture of the agent workflow.
3. Recommend the necessary n8n nodes (built-in, HTTP, function, OpenAI, etc).
4. For each node, explain its configuration and purpose.
5. Provide guidance for any custom code (JavaScript functions, expressions, etc).
6. Help me set up retry logic, error handling, and fallback steps.
7. Show me how to store and reuse data across executions (e.g. with Memory, Databases, or Google Sheets).
8. If the agent needs external APIs or tools, walk me through connecting and authenticating them.



Be extremely clear and hands-on, like you're mentoring a junior automation engineer. Provide visual explanations where possible (e.g. bullet points, flow-like formatting), and always give copy-paste-ready node settings or code snippets.



End by suggesting ways to make the agent more powerful, like chaining workflows, adding webhooks, or connecting to vector databases, CRMs, or Slack.
Read 8 tweets

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us!

Send Email!

:(