weicheng

Frontend dev: @angular & @sveltejs
(exited) |
#indiedeveloper #buildinpublic

[weicheng @weicheng_95]

I just reserved my Cloudflare Wallet tag: https://t.co/esPdCdVJZy.

Reserve yours now at https://t.co/RhluKF0YVA

[weicheng @weicheng_95]

Image generation getting faster and cheaper feels like when VPS prices dropped. Suddenly, you use it for quick mockups, not just polished outputs. The real unlock is removing the mental cost of experimenting.

https://t.co/kcjNGlUWS7

[weicheng @weicheng_95]

Each Claude release narrows the gap between what I can think and what I can delegate. I don't switch models for hype—I switch when a specific friction disappears. Sonnet 5's code improvements might be the next quiet unlock.

[weicheng @weicheng_95]

@AnthropicAI @euboid

[weicheng @weicheng_95]

I don't care about scaling laws. I care about the smallest model that can finish a task without crashing. Progress is measured in watts, not tokens.

https://t.co/PnmWhRNDgt

[weicheng @weicheng_95]

I keep seeing people dream about tomorrow’s hardware. I’m more interested in what I can build with today’s. Downloaded the weights, ran it—cool. But I’m shipping on a machine that’s already paid for.

https://t.co/FcfVZZGSIO

[Alex Finn @AlexFinn]

As a sidenote, this is a MASSIVE model

You will not be running this on a Mac Mini. In fact, you will need quite a few high end GPUs

But here's the thing: hardware will get better. Models will be made more efficient. This is the worst it will ever be

Within the next few years this will 100% be able to run on widely available consumer hardware. I'd bet ANYTHING on it

The fact that a model who's weights you can download right now is just as good as Opus 4.8 is a feat in its own

[weicheng @weicheng_95]

I keep hearing 'agentic' as the new enterprise pitch. It reminds me of 'serverless' three years ago. The biggest impact for solo devs won't be autonomous orchestration—it'll be the one cron job you stop babysitting. Focus on the simplest handoff, not the full automation fantasy.

[weicheng @weicheng_95]

I wasn't betting on open source catching up this fast—I was just hedging. GLM matching Opus changes my build logic overnight. Not because of the tech, but because the cost curve just flipped. What I run locally isn't just cheaper now; it's unrevocable.

[weicheng @weicheng_95]

I just saw Lisa Su pull out a lunchbox that runs a huge AI model locally. My monthly cloud bill for inference suddenly feels like rent I didn't need to pay. Time to rethink where my AI lives.

https://t.co/8rqbOaOTuc

[Leonard Rodman @RodmanAi]

The CEO and Chair of AMD Lisa Su walked on stage holding a lunchbox-sized PC.

Not a server.
Not a GPU cluster.

A tiny box running a 235B-parameter AI model locally.

No cloud.
No data center.
No RTX 5080.

Just one AMD chip with 128GB unified memory.

Cost: $1,400–$2,500 once.

Cloud AI: $200–$400 every month.

The shocking part?

This isn't a faster GPU.

It's the first serious argument that your AI doesn't belong in someone else's data center.

[weicheng @weicheng_95]

I first learned APIs by reading raw docs and fumbling with curl. Now the Codex plugin sets up keys, finds the right docs, and debugs. Grateful for the speed, but I miss the quiet confidence that comes from the fumbling.

https://t.co/x43eQiuwe4

[OpenAI Developers @OpenAIDevs]

Use the OpenAI Developers plugin in Codex to build faster with OpenAI tools by setting up API keys, finding the right docs, and debugging along the way.

[weicheng @weicheng_95]

I once spent a week replicating a demo of an AI that 'learns on every run.' Found it was just a clever retrain schedule. Now I smile when I see another 'self-improving' model.

https://t.co/mhJoWiVkkH

[Leonard Rodman @RodmanAi]

THE HEAD OF CLAUDE JUST SHOWED THE FUTURE OF AI IN 12 MINUTES 🤯

Fable 5 doesn't just follow instructions.

It learns, adapts, and gets better every time it runs.

[weicheng @weicheng_95]

Nvidia’s new physical AI agent workflows are impressive. Meanwhile I’m on week 3 trying to make a Raspberry Pi camera detect my cat without false positives. The real agent workflow is me, refreshing logs.

https://t.co/dnLfakedJN

[weicheng @weicheng_95]

Went through the showcase like a tourist. Missed half the details until I squinted. Same thing happens with my own product demos—people see what they expect, not what I built. The good stuff always hides in plain sight.

https://t.co/CndDDBjUDz

[OpenAI Developers @OpenAIDevs]

Look closely. There’s more in the Showcase.

[weicheng @weicheng_95]

Settings search is useful, but the real metric is how often you stay in the flow. Every config toggle is a context switch. I'd rather have fewer switches than faster navigation.

https://t.co/DVZJHio4bC

[OpenAI Developers @OpenAIDevs]

Today’s Codex quality-of-life updates start in settings.

You can now search Codex settings, with results grouped by category, so you can find what you want to change without scanning every section – this makes setup and customization easier.

[weicheng @weicheng_95]

Building with AI feels like flying only until you hit autopilot loops. The real skill is knowing when to take the controls back.

[weicheng @weicheng_95]

Hermes' desktop app kills the Telegram dependency. Now agents run natively. But the bottleneck isn't the messaging layer—it's how you specify the finish line. I track cost per session, not UI parity.

https://t.co/S4w0HvGasv

[Alex Finn @AlexFinn]

Hermes won. They just dropped their desktop app and it's excellent

It's now the best way to use AI agents on your computer

In this video I cover how to set it up, how to use it, and go through EVERY feature in the app

Bye bye Telegram

[weicheng @weicheng_95]

Automated customer discovery sounds efficient, but scanning for problems is the easy part. Understanding them deeply is where the work happens. A cron job can't replace lived experience in the niche.

https://t.co/gTKJC9a6C4

[Alex Finn @AlexFinn]

If you're not using Hermes Agent + the X API together you're a moron

Most powerful way to find money making opportunities

Here's what you need to do:

1. Sign up for the X API (developer dot x dot com)
2. Set up Hermes in telegram or desktop (see my vids)
3. Ask Hermes to set up a cron for every morning at 9am
4. Ask it to scan X for challenges people are having in your niche (describe your niche)
5. Have it give a list of the top 3 challenges, and a recommendation for how to build an app or service to solve them
6. Have it build a prototype of those apps

If you have any trouble setting up the X API just ask Hermes to do this

If you want to take this a step further, you can even have Hermes automatically build you a prototype with each cron

Do this, and I guarantee you come up with multiple business ideas every week you can make money off of

We have an unreal opportunity right now to leverage these AI tools 99% of people have no idea exist

The fact that nobody knows about them means you have an INSANE opportunity to get ahead

Now take action before someone else does

[weicheng @weicheng_95]

I've migrated setups. One-click imports are a rush, but you're quickly tied to the meter: every agent action, every background check accumulates token cost. I watch cost per session, not config portability.

https://t.co/8V7qarSBrx

[OpenAI @OpenAI]

Bring your workflow to Codex in just a few clicks.

Import settings, plugins, agents, project configuration, and more so you can keep working with fewer interruptions.

Your move.

[weicheng @weicheng_95]

GPT-5.5's Codex revenue doubled in under a week. I track absolute revenue, not launch multiples. A 2x on a small base is a rounding error.

https://t.co/0rkwYaKrpX

[OpenAI @OpenAI]

One week since the launch of GPT-5.5, and it’s already our strongest model launch yet.

API revenue is growing more than 2x faster than any prior release, while Codex doubled revenue in under seven days as enterprise demand for agentic coding tools keeps climbing.

[weicheng @weicheng_95]

A '/goal' session that builds a game for hours without intervention redefines the maker's role. You don't write code; you define intent and curate results. The limiting factor isn't the loop—it's how well you can specify the finish line.

https://t.co/aLqWAMzxkz

[Alex Finn @AlexFinn]

Pretty incredible

You have to try the new '/goal' feature in Codex

It worked for over an hour and built me an entire complex extraction shooter video game

You give it a goal, then it works endlessly until the goal is complete. It's like a Ralph loop. Can run for days

If you enable the image gen skill before you run the goal, it will even generate ALL the assets for your game autonomously. I didn't manually create ANY of the assets you see in the video

Recommendations: enable the image gen skill, put on skip all permissions, and give the prompt as much detail as you can. It will accomplish ALL of it

This has to be the sickest way to build games/ long running app tasks ever