Aldhaheri

• Technological Mind • Anti Ignorance •

[Aldhaheri @aenawi]

@monokern @grok does this paper exist or it’s fake!

[DAN KOE @thedankoe]

https://t.co/gN9cBK7feV

[Aldhaheri @aenawi]

@FishAudio Fish

[Fish Audio @FishAudio]

Today we’ve raised $52M Seed and we are announcing the public launch of S2.1 Pro.

>It can clone a voice from 5 seconds of audio
>2x faster than Cartesia & 1/6th the cost of Eleven Labs
>most expressive model with word level control over emotion, intonation, pacing etc

We support frontier AI companies including HeyGen, LiveKit, Retell, Sanas, and OpenArt all run our model in production.

If you're a business and we can't cut your voice AI costs by 50%, we'll give you 1 year of Fish Audio for free.
Book a demo: https://t.co/vHkyZf9JoG

To celebrate our first birthday, we'll give you 1 month of S2.1 Pro for free. Like, retweet, and comment “Fish” to get it.

[Aldhaheri @aenawi]

@herdrdev IM•Possible! 👍🏻

[Aldhaheri @aenawi]

Just open-sourced namux — a tiny Go CLI for brandable business + domain names.

Built for humans and AI agents: stable JSON schema, exit codes, multi-format output (table/json/csv/plain).

go install https://t.co/NsJCq4Wnp6
namux generate "coffee shop" --style brandable --ext com,io --available

Independent, MIT. Expandable backend design.

👉 https://t.co/nuh3DZYbES

#Naming #DomainNames

[Claude @claudeai]

We're extending access to Claude Fable 5 on all paid plans through July 12.

[GitHub Projects Community @GithubProjects]

Chunkr is an open-source document intelligence service that converts PDFs, PPTs, Word docs, and images into structured chunks ready for RAG and LLM pipelines.

- Layout analysis with OCR and bounding boxes
- Structured HTML and Markdown output
- Vision-language model processing
- Self-hosted via Docker Compose with configurable LLM providers

[Ayaan 🐧 @twtayaan]

Epic Games got tired of Git. So they built their own version control system. Then gave it away for free.

This is Lore. Announced at State of Unreal 2026. Open source under MIT. Available right now.

Git was never built for game development. When you are working on a project like Unreal Engine or Fortnite you are mixing source code with massive binary files. 3D models. Textures. Audio. Build outputs. Files that are gigabytes in size.

Git handles large binary files through a bolted-on add-on called Git LFS. Perforce handles them better but it needs a live server connection for everything and nobody outside your company can build tools on top of it.

Epic said none of the existing systems solved all three problems together. Large binary handling, offline work and an open specification. So they spent years building one themselves.

Here is what Lore actually does differently:

→ Works completely offline. Commits, branches, switching — all happen on your machine without touching a server.
→ Every file gets a unique fingerprint. Identical data is never stored twice across files or branches.
→ Large files are broken into chunks. Edit one part of a multi-gigabyte file and only that chunk gets re-uploaded.
→ Your machine only holds files you are actually using. Everything else gets pulled down on demand.
→ Built in Rust. Core library, server and CLI all use the same interface. No special privileged commands.
→ Official SDKs for JavaScript, Python, C# and Go ship with it.

Perforce licenses cost tens of thousands of dollars a year. Git LFS is a workaround that was never meant to be a solution.

Lore is free. Written in Rust. MIT licensed. And Epic just put it on the internet for anyone to use.

Repo here:
https://t.co/WelEpip8Ln

[Aldhaheri @aenawi]

Nobody sets out to waste money on cloud. You pick the cheapest instance, you turn off what you don't need, you read the pricing page. Then 6 months go by and the bill is 3× what you expected.

The cloud doesn't overcharge you in one big hit. It bleeds you through a hundred small cuts you never see.

Egress. Snapshots. Managed markups. Idle resources. DB add-ons. Retrieval traps.

I got tired of auditing bills manually. So I started building something to do it automatically.

@DraftMyStack — coming soon.

If you want early access: DM me or reply here. I'll add you to the waitlist.

[Aldhaheri @aenawi]

🧵 Thread #2: The Hidden Costs Nobody Shows You

Last thread I mentioned 6 things that kept inflating my cloud bills.

Today I'm explaining what each one actually is — and why they're so easy to miss.

New here? Catch up first 👇
https://t.co/RhsshwhBeC

[Aldhaheri @aenawi]

🧵 I wanted to build AI apps. Turns out coding was the easy part.

Here's what nobody tells you about hosting costs — the traps I kept stepping in before building @DraftMyStack.

I watched the ChatGPT boom from the sidelines — 2022, 2023. Impressive, sure. But I'd been writing code long enough to know a demo isn't a product.

You need it secure. You need it fast. Anyone can spew out spaghetti in a weekend — making it production ready takes real work.

Then the AI coding tools started showing up. Codex. Gemini. Lovable. Claude was the one that clicked for me — it actually remembered context across a whole conversation, not just whatever I said last. That changed things. Suddenly vibe coding wasn't a party trick. It was something I could use to move faster without cutting corners.

[Aldhaheri @aenawi]

6. Storage tiers & retrieval

"Archive this to Glacier, it's basically free."

S3 Glacier Deep Archive: $0.00099/GB/month. 500GB of old logs? $0.50/month. Almost nothing.

The catch is retrieval. When you actually need that data back: $0.02/GB + $0.05 per 1,000 requests. That 500GB archive now costs $10+ to restore. And if you need it fast (expedited retrieval), it's 5× more.

So you archived 500GB to save $11/month on standard storage. Six months later you need it back. Retrieval cost: $10+. Six months of "savings": $6.50.

Archival isn't a shortcut. It's a bet that you won't need the data. If you lose that bet, you pay.

[Min Choi @minchoi]

We are cooked.

China's Alibaba just revealed Wan Streamer.

AI agents can now see you, hear you, and talk back on video in real time.

This is not voice mode anymore 🤯

[Ole Lehmann @itsolelehmann]

i'm 100% convinced this is the future of espionage

china's national university of defense technology just built a drone the size of a real mosquito

packed into that tiny body are a micro camera, a microphone, and sensors that can pick up images, sound, and even electronic signals

the wings flap like an insect instead of spinning like a normal drone propeller, so it stays nearly silent and moves like something alive

you fly it from your phone over bluetooth

it's under 0.3 grams, about 2cm long, with three legs as thin as a human hair

so it's basically a flying bug that can sit on a windowsill and record the whole room.

you'd swat it on reflex. the idea that it's recording you never crosses your mind

the use-cases here are infinite:

> surveillance that walks past every metal detector and camera because it reads as an insect

> tapping a room nobody would ever think to sweep for bugs

> battlefield recon where a soldier sends one ahead to map a building before anyone steps inside

the catch right now is power.

even with the best batteries it flies for maybe 1-3 minutes, it's fragile in wind, and it can only carry the most basic sensors.

it's also still a lab prototype

but this is the worst this tech will ever be

the moment battery density catches up, you won't be able to trust a mosquito in your room again

[Aldhaheri @aenawi]

Not live yet.

If you want early access when the waitlist opens → follow @aenawi and @DraftMyStack. I'll announce it here.

No hype. Just numbers when it's ready.

If you've ever:
• Opened a cloud bill and it was way higher than you expected
• Lost hours comparing pricing pages
• Felt like the pricing was designed to confuse you

Follow along. I'm posting everything I learn.

[Aldhaheri @aenawi]

🧵 I wanted to build AI apps. Turns out coding was the easy part.

Here's what nobody tells you about hosting costs — the traps I kept stepping in before building @DraftMyStack.

I watched the ChatGPT boom from the sidelines — 2022, 2023. Impressive, sure. But I'd been writing code long enough to know a demo isn't a product.

You need it secure. You need it fast. Anyone can spew out spaghetti in a weekend — making it production ready takes real work.

Then the AI coding tools started showing up. Codex. Gemini. Lovable. Claude was the one that clicked for me — it actually remembered context across a whole conversation, not just whatever I said last. That changed things. Suddenly vibe coding wasn't a party trick. It was something I could use to move faster without cutting corners.

[Aldhaheri @aenawi]

So I started building what I couldn't find.

@DraftMyStack.

Compare providers honestly. See all the costs, not just the ones on the landing page. Clear numbers, no math degree needed. Scales from solo devs to teams.

I saw the trap coming and I'm building a way out.

For myself first. Hopefully for you too.