Yashica Jain

Founder & CEO @ Automatez AI | We build AI Automation systems, I advise businesses 1:1, and share practical AI builds on YouTube

[Yashica Jain @yashicajain_]

Free 28-minute course on mastering GrokBot, now that it costs $20/mo instead of $300/mo.

From one prompt to a fully autonomous team of agents with their own computers, inboxes and schedules:

@xai @cursor_ai #grokbot

[Yashica Jain @yashicajain_]

@Ravi26329535 @xai @cursor_ai Yes, it's already uploaded there!

[Yashica Jain @yashicajain_]

Free 28-minute course on mastering GrokBot, now that it costs $20/mo instead of $300/mo.

From one prompt to a fully autonomous team of agents with their own computers, inboxes and schedules:

@xai @cursor_ai #grokbot

[Yashica Jain @yashicajain_]

Watch the Full Video here: https://t.co/k0ZbaOjcyW

[Yashica Jain @yashicajain_]

Finally, an AI video model that doesn't change the original footage.

We're at a stage now where you can upload a video and ask it to make changes:

- change the location
- change the character
- change the objects
and more

All without losing the feel of the original video. And I find that genuinely interesting.

I've made a full tutorial on this new AI video-to-video model on my channel.

Link in the comments. 👇

[Yashica Jain @yashicajain_]

Step-by-Step Tutorial on Youtube: https://t.co/Z28m6rXnBf

[Yashica Jain @yashicajain_]

Your AI voice agent isn't taking anyone's job.

It's picking up the work that was already falling on the floor.

Someone enquires at 11:40pm on a Sunday.

Who exactly was calling them back? Not your team, they're asleep. Not Monday morning either, that's spent on the pipeline that already exists. Realistically that lead gets a call Tuesday afternoon, if anyone remembers.

That follow-up wasn't being done badly. It wasn't being done at all.

This is the part the replacement argument keeps missing.

Almost every business I've built one of these for had the same gap, and none of them had a person in it. There was no SDR losing a seat. There was a spreadsheet of enquiries nobody had the hours to work through, quietly aging out.

You cannot replace a job that was never staffed.

What actually gets replaced is the batch. The Tuesday calling block where someone works down a list that went cold on Sunday night. That ritual dies, and nobody should mourn it, least of all the person doing it.

Here's the part people find harder to accept.

The reason it works isn't that it's clever. It's that it's available.

It calls at 11:41pm on a Sunday. It calls again at 6am. It makes call 40 sound like call 1, because it has no memory of the last 39 people who hung up. At roughly 5 cents a minute, a second attempt stops being a decision anyone has to justify.

None of that is intelligence. It's just presence.

And presence was always the bottleneck. Not talent, not scripts, not CRM hygiene. Somebody being awake and on the phone in the five minutes where the lead still cares.

What it does not replace: the actual conversation. Someone still has to run the demo, read the room, and close. This books the appointment. It doesn't do the job that appointment exists for.

The call in this video is a real one, uncut. Watch the order it works in. It says who it is, checks it has the right person, asks if now is okay, explains why it's calling, and confirms the interest is still live. Only then does it pitch.

Most human SDRs skip half of that.

Full step-by-step guide on YouTube. Link in Comments.

If you'd rather not build it yourself, I set these up for businesses. Send me a DM and tell me what your follow-up currently looks like.

[Yashica Jain @yashicajain_]

Well... Anthropic extended Fable access again.

Fable 5 is now available on all paid plans through July 19, and Claude Code's weekly limits are staying 50% higher too.

That tells me one thing. People are using it.

Companies don't keep extending previews unless they're seeing strong engagement.

But I also wonder...

Is this Anthropic's response to the growing competition from GPT-5.6?

I've spent the last week building with Fable, and three things stood out:

• It's the best model I've used for turning messy ideas into a clean implementation plan.
• Frontend quality is consistently excellent. It produces polished, cohesive UIs with surprisingly little guidance.
• It handles large project context better than most models I've tried, which means less repeating myself.

The downside?

I wouldn't use it for every task.

My workflow is simple:

Fable for planning, architecture, and complex reasoning.
A cheaper model for repetitive implementation.

That hybrid workflow feels like the sweet spot right now.

What's your take?

#Anthropic #OpenAI

[Claude @claudeai]

We're extending Claude Fable 5 access on all paid plans, as well as keeping Claude Code’s weekly rate limits 50% higher, through July 19.

[Yashica Jain @yashicajain_]

link to the yt video: https://t.co/DuCeuHoGcj

[Yashica Jain @yashicajain_]

Hermes 'Cloud' might be the easiest way to run a real 24/7 AI agent RIGHT NOW.

No self hosting. No "my agent sleeps when my laptop does."

I connected it to Telegram in under 60 seconds.

Gave it a task. Gave it feedback.
Then it saved my feedback as a skill. It will not get it wrong twice.

That is the whole point. Persistent memory is why Hermes keeps getting better while every other agent starts from zero.

I covered how cloud differs from self hosting and desktop, the pricing, the full setup, and the use cases.

Watch it here👇🏻 or on my youtube channel!

#hermesagent #aiagents

[Yashica Jain @yashicajain_]

OpenAI launches ChatGPT Work.
Claude resets rate limits for everyone.

The competition is getting intense.

Meanwhile, we’re just sitting here enjoying better models, higher limits, and more reasons to build 😂

Please continue.

[Yashica Jain @yashicajain_]

Watch the Full Breakdown on Youtube: https://t.co/y4qLivP5wV

[Yashica Jain @yashicajain_]

You're probably overpaying for AI coding.

Same prompt, one shot: GLM 5.2 (½ the price) vs Opus 4.8.

One pulled ahead, and not the one I expected.

Which would you pick? 👇

Full breakdown on YouTube 🔗

[Yashica Jain @yashicajain_]

And these are just 5 of them. There are a lot more in the full breakdown.

🎥 Full video: https://t.co/umfxNXxHn3

What's your best token-saving habit?

[Yashica Jain @yashicajain_]

Most people think their messages are what burns through their Claude limit.
That's not it. There's more to it.

Here's how to stop hitting the wall right before the good part 🧵

[Yashica Jain @yashicajain_]

5. Build once, run it free.

Need to check a site daily? Don't make an LLM do it on repeat.

Have Claude Code write the script once. It runs forever without touching your limit.