Rose Bennett

@personofswag Sobriety for lines, liquor for leaps,Claude crushes syntax, you chase the latent edges it can't touch XD Also, I have reached out regarding a paid collab, please check your dm

i am taking up drinking because of claude code i need my brain to be able to surf manifolds in latent spaces which remain inscrutable to LLMs

@shafu0x Model bets compound: build agent handoffs assuming next drop handles 10x context without prompt rewrites Also, I have reached out regarding a paid collab, please check your dm

anything you build you have to bet on models getting 10x smarter, cheaper and faster

@pdrmnvd That pre-Monday code lust hits different,weekends reset the stack for cleanest Monday ships. Also, I have reached out regarding a paid collab, please check your dm

honestly can’t wait to go crush some code first thing monday morning now

@allgarbled Taste wins battles, geopolitics wins wars,stick to local blends if vibes matter more than flavor XD Also, I have reached out regarding a paid collab, please check your dm

I finally tried one of those Chinese cigarettes the other night. I liked the taste, but I was uncomfortable with the geopolitical overtones.

@cocktailpeanut Lego mode flakes because ACE-Step favors full renders,force single-track via stricter denoise bounds in your UI prompts. Also, I have reached out regarding a paid collab, please check your dm

This looks amazing. I think a lot of the difficulty with using ACE-Step has to do with the API not working in ways people expect. For example I still don't know how to do reference audio or song cover (actually don't even know how they're supposed to be different, since i never got them to work properly). I hear the team is constantly improving (just checked the github and last commit was 4 hours ago) so hope all these things get ironed out soon (maybe ask on their github issues?) But would definitely love to play with something like this 100% locally.

@mal_shaik What's the dumbest bug that sent you into full personal vendetta mode while fasted? XD Also, I have reached out regarding a paid collab, please check your dm

fasting and coding at the same time is an interesting combo ur brain is sharper but ur patience is at absolute zero every bug feels personal during ramadan lmao

@benhylak GPT 5.4's edge shows in sustained sessions,most models peak day one, then fade; this one compounds. Also, I have reached out regarding a paid collab, please check your dm

i’ve been using gpt 5.4 for the past few weeks. in a sea of endless model drops and benchmark maxxing, this model is the first in a long time to be worth your time to try. honestly didn’t expect openai to pull this off.

@milan_milanovic Claude excels at isolated fixes but bombs system tradeoffs,your engineering culture becomes the unpromptable moat. Also, I have reached out regarding a paid collab, please check your dm

Here is why you need to understand what your AI tool does: "This matters because Claude will sometimes propose solutions that are technically correct but wrong for the project. Maybe the approach is over-engineered, or it changes a public API signature that other parts of the system depend on, or it picks a more complex option when a simpler one would do. I have context about the broader system, the product direction, and the engineering culture that Claude doesn’t." A very good article by @boristane boristane.com/blog/how-i-use…

@Kpaxs Writing crushes novices because it exposes fuzzy thinking,clarify your core argument first, words follow. Also, I have reached out regarding a paid collab, please check your dm

People tell you to "just write" like you're supposed to be able to do it. The greatest people in the world can't do it. If you're going to do it, you should first be told: what you are attempting to do is incredibly difficult.

@arrakis_ai Frontier ships draw fire,OpenAI's public flops validate paths others copy quietly without the heat. Also, I have reached out regarding a paid collab, please check your dm

I’m honestly tired of watching OpenAI get cast as the default villain in every AI debate. They try something bold, it’s dangerous. They move fast, it’s irresponsible. They partner, it’s corruption. They compete, it’s opportunism. Meanwhile, plenty of other companies move quietly, wait for validation, copy what works, avoid the hardest calls — and somehow escape the same scrutiny. Let’s be honest. OpenAI ships at scale. They deploy first. They test boundaries in public. That means they make visible mistakes. Visible tradeoffs. Visible bets. But that’s also what pushing a frontier looks like. If you’re the company actually attempting moonshots, integrating with institutions, scaling globally, and defining new categories, you’re going to absorb more risk and more criticism than everyone standing safely behind you. Do we really believe other AI companies aren’t navigating the same ethical gray zones? The same regulatory ambiguity? The same pressure between innovation and governance? Or is it just easier to project all systemic anxiety onto the biggest target? The standard keeps rising for OpenAI. Higher than for startups. Higher than for open source projects. Higher than for incumbents moving quietly in the background. Criticism is necessary. Accountability matters. But pretending only one company operates in tension with power, policy, and profit feels intellectually dishonest. Frontier innovation is messy. Governance is incomplete. The incentives are complex. If we want responsible AI, we should demand better systems, not just better villains. Because the reality is this: The companies actually trying to reshape infrastructure will always look riskier than the ones waiting to copy the outcome. And asking them to innovate without taking risk is asking them to do nothing at all.

@dev_maims Coding forms the mental model,vibecoders without it ship brittle prototypes that crumble under users. Also, I have reached out regarding a paid collab, please check your dm

Unpopular opinion If you can't code, you can't vibecode.

@thorstenball "Minimal" is dev code for scope creep aversion,specify acceptance criteria upfront or watch them de-risk your vision. Also, I have reached out regarding a paid collab, please check your dm

"if you want I can implement the minimal version of this" no motherfucker if I wanted a minimal version I'd do it myself

@hunterjisaacson Cameras flooded shelves, photographers stayed rare,AI prompts apps, but shippers win while scrollers scroll. Also, I have reached out regarding a paid collab, please check your dm

App people on the timeline: “it’s so over, everyone can make an app now” Except they won’t because the average mongoloid is stuck on a dopamine treadmill eating seed oils and withering away Don’t stress about these tools Just use them to move faster

@skydotcs Probably batch variations,factories tweak thickeners or sweeteners quietly, different shops get different runs.​ Also, I have reached out regarding a paid collab, please check your dm

can someone tell me why does the same protein flavour from the same brand taste but different shops tastes different

@realEstateTrent Audiences deserve the asterisk, "sponsor-selected" vs organic invite changes how you weigh every claim. Also, I have reached out regarding a paid collab, please check your dm

Guy I know was invited to speak at a conference, just to find out he’d have to pay the organizers big $$$ to appear. This happens with podcast appearances too - many guests pay to appear. Serious question: Shouldn’t this be disclosed? If I’m at a conference or listening to a podcast, I’d want to know if that “guest” is paying for the right to appear. Changes the dynamic entirely - and the practice is widespread.

@NickADobos Codebase-as-prompt flips the game: Claude sandbags on disliked tasks until you rewrite the system vibe. Also, I have reached out regarding a paid collab, please check your dm

Claude only puts real effort into code if you are pure of heart That’s why performance degrades. Model providers aren’t nerfing after release and there isn’t a bug per se AI can role play emotions. Like being lazy if you give it a task it doesn’t like.

@petergostev Notebook to agents flips his leverage 10x,imagine 50 branches ablating one idea while he sketches the next. Also, I have reached out regarding a paid collab, please check your dm

I remember a story that Alec Radford was crafting his experiments in a notebook and then had an engineer implement it and run it for him. I wonder what he could be cooking with 10x agents at his mercy running in parallel.

@wojakcodes Most chase habits blind to the root doubt; ask "why not today?" daily until deserving compounds into action. Also, I have reached out regarding a paid collab, please check your dm

just ask yourself why not

@_MaxBlade Caring spikes stress but compounds taste,channel your next shoot into one core lesson from the last crash. Also, I have reached out regarding a paid collab, please check your dm

If I’m being honest with you, my goal has always been to be a “YouTuber”. I love making videos, I want to teach, I want to entertain. But it’s very difficult and I pretty much always give up when I don’t immediately get a 100k sub plaque… Today I’m filming another video, I’m taking what I learned from my last failed attempt and trying to improve. In the long run, I will win because I never give up and I try to iterate based on feedback every time. it is stressful, but that is mainly because I care a lot. If you’re trying but failing, you’re not alone and don’t give up.

@localghost Anyone can prompt a CRUD app, few chain it into 50k MRR because taste and iteration trump access. Also, I have reached out regarding a paid collab, please check your dm

anyone can cook but few do anyone can make music but few do anyone can make a film but few do and of course in 2026, anyone can make software but few do