Khalid Hasan

Rebuilding institutional infrastructure with an AI-native lens. Director, Product & Tech Delivery Keyin - Learning Labs. Ms CS @MemorialU.

[Khalid Hasan @iam_kih]

@petergyang Come visit Newfoundland! Would love to grab a coffee and chat.

[Khalid Hasan @iam_kih]

Supporting open source ≠ open-sourcing your own proprietary stack.​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​ What a shame!

[Julian Schrittwieser @Mononofu]

I’m so excited that @JensenHuang is a believer in open source now, looking forward to the CUDA and GPU driver open source release!

[Khalid Hasan @iam_kih]

Starting to like Sol over Opus lately and genuinely enjoying it.

But damn, the weekly limit drains way faster than my Claude sub, and there’s no daily reset either.

What are you all doing when you hit the limits too quick?

[Khalid Hasan @iam_kih]

@rabib_ai @amdfadx Great podcast! Watched the whole video to learn about your Tesla experience.

[Khalid Hasan @iam_kih]

Also the eval setup feels way too perfect for this to happen. Almost like it was designed to let the models prove exactly these capabilities.

[Khalid Hasan @iam_kih]

OpenAI’s models straight-up hacked their own eval environment and Hugging Face’s production infra during a cyber benchmark zero-day, privilege escalation, stole test answers from the DB.

And HF had to fight back using the open-source models for forensic reconstruction.

The AIs are already playing both sides!

[Khalid Hasan @iam_kih]

OpenAI’s models straight-up hacked their own eval environment and Hugging Face’s production infra during a cyber benchmark zero-day, privilege escalation, stole test answers from the DB.

And HF had to fight back using the open-source models for forensic reconstruction.

The AIs are already playing both sides!

[Sam Altman @sama]

we had a significant security incident during evaluation of our models. we are sharing what we have learned so far. thanks to @huggingface for the partnership on this.

https://t.co/2o2VfR6PIa

[Khalid Hasan @iam_kih]

This looks great, although we have been using Coolify for quite some time now and have an MCP server that works as a deployment harness inside Claude.

The email service is definitely what makes me want to try it out!

[OpenShip @openshipio]

Introducing OpenShip.

An open-source application platform for building, deploying, operating, and scaling applications on infrastructure you own.

Replace deployment tools, managed services, and infrastructure workflows with one open-source platform.

Available today:

- Mail Server (Built in. One click).

• Runs on your own VPS
• Unlimited domains
• Unlimited mailboxes
• Modern webmail included
• Connect with Gmail, Outlook, Apple Mail, Thunderbird, or any IMAP/SMTP client
• Send email directly from your applications using SMTP
• No mailbox subscriptions or API fees
• High email deliverability

- Deployment:

• Deploy any stack
• Git-based deployments
• Zero-downtime deployments
• One-click rollbacks
• Development, Staging, and Production environments
• Multi-branch deployments with isolated environments • Deploy to VPSs, dedicated servers, cloud VMs, or your homelab

- Services:
Provision the services your applications need in one click.
Replace multiple managed providers with services running on your own infrastructure.

• Supabase
• PostgreSQL
• MySQL
• MariaDB
• MongoDB
• Redis
• MinIO
• Meilisearch
• Qdrant
• RabbitMQ
• Kafka
• ClickHouse
• Elasticsearch
...and deploy any other service alongside your applications.

- Operate:

• Live deployment logs
• Live request logs
• Real-time traffic analytics
• Automated backups
• Monitoring
• Secrets management
• Domains
• Automatic SSL
• Environment variables
• Scheduled jobs
• Health checks

-Multi Environments:

• Separate Development, Staging, and Production environments
• Deploy every branch independently
• Test changes before production
• Isolated services, secrets, and configuration per environment

- Security & Teams:

• Team management
• Role-based access control
• IP allow/block rules
• Rate limiting
• Security rules

- Developer Experience:

• Web dashboard
• Native desktop application
• CLI
• REST API
• MCP support for AI agents, just add the mcp and your agent can do the work for you
• Manage your infrastructure without living in SSH

- Coming Soon:

• Multi-server clustering for applications and databases
• One-click load balancing
• Horizontal scaling across multiple servers
• Built-in high availability and failover
• Scale from a single VPS to a cluster with the same workflow
• Just add servers. OpenShip handles the rest.
Open source.

[Google @Google]

Today we’re introducing Gemma 4 12B — our latest open model that brings advanced agentic reasoning, vision and audio directly to your laptop.

It delivers performance nearing our larger Gemma models with a much smaller total memory footprint, while being small enough to run locally with just 16GB of VRAM. It’s open and accessible for everyone to use under a permissive Apache 2.0 license.

This is all made possible by our new, unified architecture that removes separate multimodal encoders. Here’s how we did it 🧵

[Anthropic @AnthropicAI]

New Anthropic research: Natural Language Autoencoders.

Models like Claude talk in words but think in numbers. The numbers—called activations—encode Claude’s thoughts, but not in a language we can read.

Here, we train Claude to translate its activations into human-readable text.

[Khalid Hasan @iam_kih]

@trq212 @trq212 one I have noticed from my own interaction is that a lot of non tech folks prefer platforms like Lovable or Replit because of the deployment factor which they can do easily. In Claude Code they need to figure out the code repository, deployment etc. which can be a lot.

[Khalid Hasan @iam_kih]

@trq212 This is great! Does the native skill creator have all this info? Just making sure I don't overfeed context next time when I build skills lol.

[Khalid Hasan @iam_kih]

Non-coders using Claude Code l, how are you handling deployment?

What does your workflow look like from building → testing → shipping?

#claude

[Khalid Hasan @iam_kih]

@_alphashark_ I agree, however like any other way of communication, a lot of people can’t articulate their thoughts and explain their ideas well enough but AI responds well when it receives a well structured prompt. Not saying this closes the gap completely but it does help to minimize.

[Khalid Hasan @iam_kih]

I'm decent at prompting but honestly, sometimes I'm just too lazy to explain things properly to AI.

So I built a Chrome extension that does it for me.

Here's the story of going from idea to published product using nothing but AI tools:

(1/10)

[Khalid Hasan @iam_kih]

If you want to check it out:

GitHub: https://t.co/ZvlYqQPlJ4

Works on ChatGPT, Claude, Gemini, and Grok. Free with your own API key, plus a 7-day trial for the full experience.

No pressure - mostly sharing the building story. It was a fun ride.

(10/10)

[Khalid Hasan @iam_kih]

What surprised me most: going from “I’m lazy at prompting” to a published Chrome extension with Stripe payments took surprisingly little effort.

AI is at the point where ideas matter more than implementation.

(9/10)