←
Luke2,203 posts
Luke avatar

Luke

@ImLukeF

heyparley.app My personal askfrankie.com.au Releasing Soon reveloplus.com Win Real Money! getrefunded.com.au

◷ Joined August 2025▧ 455 Photos and videos
Luke avatar
Luke@ImLukeF·48m

You know what @GoogleAI should do. Let me use my subscription in other apps and services even ones I build. Just a thought.

Luke avatar
Luke@ImLukeF·1h

whats with all the random model released lately?

Gemini 4 Neon is rolling out now. Sitting between Carbon and Hydrogen, it brings frontier intelligence and fast task completion for complex software engineering, knowledge work, and science.

Post attachment
Luke avatar
Luke@ImLukeF·10h

The past 2 weeks have been a real eye-opener for me, and I think software is dead, but at the same time, this is the most important time to learn AI and SWE. I think traditional software companies like @Adobe & @Apple & @claudeai should adapt and open the gates because AI will change the world, and software as we know it will change forever.

Meet Remy! He's my AI agent built from the ground up. (Yes, I took design inspiration from other agents), but I've built a better, cheaper, smarter system, taking the best bits from the big labs (mainly GrokBot and Muse). It's far cheaper for me to run, I can run it all day, every day, for under $50 a month. It's smarter and runs on Cloudflare agents and my VPS, which is more powerful and has access to a GPU server. I'll share more in the coming days. Over the last two weeks, I've cancelled or cut subscriptions and built my own in-house software: better, cheaper, faster, and always updatable. More to come!

Luke avatar
Luke@ImLukeF·10h

Meet Remy! He's my AI agent built from the ground up. (Yes, I took design inspiration from other agents), but I've built a better, cheaper, smarter system, taking the best bits from the big labs (mainly GrokBot and Muse). It's far cheaper for me to run, I can run it all day, every day, for under $50 a month. It's smarter and runs on Cloudflare agents and my VPS, which is more powerful and has access to a GPU server. I'll share more in the coming days. Over the last two weeks, I've cancelled or cut subscriptions and built my own in-house software: better, cheaper, faster, and always updatable. More to come!

Hey dw videos will get better im learnign how to prompt still 😂😉

0:12
Luke avatar
Luke@ImLukeF·17h

Tell me Adobe isn't such a scumbag company. I installed 1 app. Adobe Acrobat. @Adobe then installed 99 additional things. I have now cancelled every Adobe service and subscrition and will never be using them ever again. I suggest you do the same.

Post attachment
Luke avatar
Luke@ImLukeF·Nov 17

Bro is this the worst launch and model release in the history of AI? Not a single good thing has been said so far about either of these models. And I guarantee ultra fast on 200 will kill your usage in 5mins. If Astra usage doesn’t last at least a week I’m sorry to say this might be the end of OpenAI for now.

OpenAI@OpenAINov 17

Some of you have been asking very politely. Some of you have just been asking. GPT-6.2 Astra and Sol are now available. Super efficient and built to perform where it counts—the real world. openai.com/index/introducing…

1:59
Luke avatar
Luke@ImLukeF·Nov 17

I’ve built my own agents now in heyparley.app but let’s give muse a quick run.

Post attachment

🚨 MUSE LAUNCH ALERT 🚨 g’day, australia!! muse just arrived down under 🇦🇺 just make sure to keep it safe from all those wild critters

0:20
Luke avatar
Luke@ImLukeF·Nov 17

Damn man. I really want to like a GPT model. But it just ain’t happening this release I guess.

Theo - t3.gg @theo Nov 17 Why I hate GPT-6.2 Astra During early testing, I found myself so frustrated with this model that I spent a day writing a letter to OpenAI about it. They fixed a small handful of my complaints, but not enough for me to

Luke avatar
Luke@ImLukeF·Nov 17

Umm, someone help! I chewed through a whole Pro 200 sub in 2 days, and all I use is SOL. Someone needs to find @theo and take T3 Code the fuck away from me. I literally have not stopped. You can literally see when I started using it. Although it makes me wonder if it's not as efficient. I doubt it, but you never know. Either way, enjoying my time.

Post attachmentPost attachmentPost attachment
Luke avatar
Luke@ImLukeF·Nov 17

This is not sustainable, which is why we've been seeing many usage downgrades lately. AI needs to get much cheaper, but I don't see how the labs could match our monthly sub with current usage at API prices. Thoughts?

Post attachment
Luke avatar
Luke@ImLukeF·Nov 16

nah but why is my token usage gone through the roof the past 7 days. T3 Code has been a game changer. Thanks @theo & @jullerino. I still use the Codex app for computer use that's it. Fable 5.2 has been insanely good, I find Opus 5.5 just as good. GPT better do something fast.

Post attachmentPost attachment
Luke avatar
Luke@ImLukeF·Nov 16

I know its been mentioned many times but Gemini Models are very slow.

Post attachment
Luke avatar
Luke@ImLukeF·Nov 13

Got Gemini 4 Hydrogen to build me a WebGPU physics engine in the browser. 500 brick wall, 1,000 body pile, still holding 120fps. Game devs should be a little nervous lol

0:43
Luke avatar
Luke@ImLukeF·Nov 13

Shout out @opencode for being able to accept a Google API key and just display the current models. Fuck @antigravity & @GoogleDeepMind for making your products so incredibly hard to use; it's just not worth it. (Not like the models are good anyway)

Post attachment