Md Kashif Ali

Discovering how tech impacts us daily 🚀
Business Enquiries: [email protected]

[Md Kashif Ali @techfireco]

Is this true?

[Md Kashif Ali @techfireco]

@rdbuilds7 https://t.co/BgjARCLvsm

[Md Kashif Ali @techfireco]

Someone put Fable 5 on the pirate bay, 3.4TB

[Md Kashif Ali @techfireco]

Antigravity just cooked every bloated AI interface.

The Antigravity CLI dropped today. It brings the full agentic power of Antigravity 2.0 - multi-step reasoning, tool calling, and multi-file editing - directly to your terminal.

It’s written in Go, ultra-snappy, and runs on Gemini 3.5 Flash.

What it actually does:

→ Delegates parallel background tasks to autonomous subagents
→ Executes complex multi-file code editing with zero keyboard lag
→ Authenticates remote SSH sessions silently via keyring sign-in
→ Syncs settings and core preferences instantly with the desktop app

It uses the exact same core agent harness and app server as the flagship product. You get elite orchestration and fast-path approvals without ever lifting your hands off the keyboard.

Heavy desktop software isn't the tax for agentic power anymore. The terminal is.

Want the full installation blueprint + SSH setup guide? Like + comment "CLI" and I’ll DM it to you for free (must be following).

[Md Kashif Ali @techfireco]

@GeminiApp 3.5 Flash looks promising.

[Md Kashif Ali @techfireco]

This is created by @GeminiApp Omini

Say Goodbye to @heygen & @ElevenLabs

[Md Kashif Ali @techfireco]

X's Secret Algorithm is Now Public And It Explains Why Your Posts Get 0 Views

[Md Kashif Ali @techfireco]

https://t.co/GMp8rh3tfx

[Md Kashif Ali @techfireco]

https://t.co/GMp8rh3tfx

[Md Kashif Ali @techfireco]

@Phonenurd @geekyranjit thats why I have seen MKBHD had subscribed him in 2014,

[Md Kashif Ali @techfireco]

People still paying for Claude meanwhile me after reading this article:

[Md Kashif Ali @techfireco]

https://t.co/NnjkOJylIE

[Md Kashif Ali @techfireco]

https://t.co/NnjkOJylIE

[Md Kashif Ali @techfireco]



[Md Kashif Ali @techfireco]

Everyone is wasting their compute on standard transformer models.

SubQ just launched the first fully sub-quadratic LLM architecture, and it's a massive shift.

It reasons across 12M tokens (entire repos, 6 months of PRs) at 150 tokens per second.

I'm using it as the long-context layer for Claude Code to map codebases 10x faster.

Stop paying for compute you don't need. Bookmark this to upgrade your agent stack.

[Alexander Whedon @alex_whedon]

Introducing SubQ - a major breakthrough in LLM intelligence.

It is the first model built on a fully sub-quadratic sparse-attention architecture (SSA),

And the first frontier model with a 12 million token context window which is:

- 52x faster than FlashAttention at 1MM tokens
- Less than 5% the cost of Opus

Transformer-based LLMs waste compute by processing every possible relationship between words (standard attention).

Only a small fraction actually matter.

@subquadratic finds and focuses only on the ones that do.

That's nearly 1,000x less compute and a new way for LLMs to scale.

[Md Kashif Ali @techfireco]

What if Govt of India designed YouTube?

[Md Kashif Ali @techfireco]

LinkedIn is really trying to sell Premium.

[Md Kashif Ali @techfireco]

Job search tools charge ₹5,000/month to people who are already jobless. 🤯

One developer said no. Built an AI that evaluated 740 jobs, generated 100+ tailored CVs,
and landed him Head of Applied AI.
Then open-sourced the whole thing. Free. Local. No subscription.

Save this this is your unfair advantage.

Companies use AI to filter you out. Now you can filter them. 👇

[Md Kashif Ali @techfireco]

@OpenAI prompt? 😂

[Md Kashif Ali @techfireco]

Hey @googleflowmusic your webpage not live, make it

[Md Kashif Ali @techfireco]

What monthly subscription is 100% worth its cost?

[Md Kashif Ali @techfireco]

Someone reverse-engineered how Claude actually thinks and open-sourced it. 🧠

Not a bigger model. Not more parameters.
Just the same small set of layers... looped 16 times per thought.
770M parameters matching a 1.3B model. That's the theory.

Save this, this changes how you think about AI forever.

Does size even matter anymore? Comment below 👇