ทวีตที่ปักหมุด
Who remembers this amazing @threejs music video from days past?
@elliegoulding @HelloEnjoy
helloenjoy.itch.io/lights
English
Aaron Layton
2.8K posts
Aaron Layton
@AaronLayton
Nerd? Hell yeah! Sleep’s a myth
Nottinghamshire, UK เข้าร่วม Mayıs 2009
6.4K กำลังติดตาม11.1K ผู้ติดตาม
@nickbakeddesign AI got you 80% there and you was probably just 1 prompt away from making it look like the design on the right but reached for familiarity 😁
English
@johnholowach @kimmonismus Anyone who says the face on the right is not objectively better than the one on the left clearly are just trying to rage bait people.
If a game came out today that had the graphics on the right, every gamer would swoon over it saying what a great job they did.
English
Important update: at the press conference I was able to talk to NVIDIA employees about DLSS5.
They assured me that DLSS5 makes no changes to the characters at all (!), no changes to the assets, and no "AI filter" is applied to the characters. It's solely the lighting that makes it appear that way; they are completely the same, unaltered assets.
Chubby♨️@kimmonismus
I am seriously so annoyed by all those AI haters. DLSS5 looks phenomenal; people hate on it cause „it’s AI“. That’s the sole reason.
English
@sohrabhossain39 Would have been better if you posted images with each of these prompts 😏
English
@JudgementKazzer What's all the fuss? It looks great, yeah it's adding detail that's not their before but isn't that the point?
Even if you say it's changing up the look and feel, I am sure they will have a toggle for certain bits. Let the user decide ¯\_(ツ)_/¯
English
Why is everyone upset it can't look that ba- SWEET MOTHER OF GOD WHAT THE FUCK
NVIDIA GeForce@NVIDIAGeForce
Announcing NVIDIA DLSS 5, an AI-powered breakthrough in visual fidelity for games, coming this fall. DLSS 5 infuses pixels with photorealistic lighting and materials, bridging the gap between rendering and reality. Learn More → nvidia.com/en-us/geforce/…
English
@ipwanciu It can break and a lot if you carelessly add it. If pages have different width, or landing/others have different heading positioning etc - it will make terrible experience and may look bad in certain cases in Chrome or Safari specifically (each has own quirks).
English
@UnslothAI @Alibaba_Qwen Nice, I was able to run up and test the 9b model with a 128k Context length on a 16gb 4080 Super and it worked great
ollama launch --model qwen3.5:9b
I was able to do tool calls and spin up multiple agents with Claude Code
English
Qwen releases 4 new Qwen3.5 Small models!
Qwen3.5: 0.8B • 2B • 4B • 9B
Run Qwen3.5-0.8B, 2B and 4B on your phone. Run 9B on 6GB RAM.
The vision reasoning LLMs perform better than models 4x their size.
GGUFs: huggingface.co/collections/un…
Guide: unsloth.ai/docs/models/qw…
Qwen@Alibaba_Qwen
🚀 Introducing the Qwen 3.5 Small Model Series Qwen3.5-0.8B · Qwen3.5-2B · Qwen3.5-4B · Qwen3.5-9B ✨ More intelligence, less compute. These small models are built on the same Qwen3.5 foundation — native multimodal, improved architecture, scaled RL: • 0.8B / 2B → tiny, fast, great for edge device • 4B → a surprisingly strong multimodal base for lightweight agents • 9B → compact, but already closing the gap with much larger models And yes — we’re also releasing the Base models as well. We hope this better supports research, experimentation, and real-world industrial innovation. Hugging Face: huggingface.co/collections/Qw… ModelScope: modelscope.cn/collections/Qw…
English
@zuess05 Marketing is hard if you are a developer at heart. You have to fundamentally shift focus to really sink into it.
I guess that is why only a small % of us actually end up making a product that works, not just a side project
English
Everybody says "spend 50% of your time marketing." 🗣️
But absolutely nobody tells you how to actually market.
Reddit bans you for self-promo,
Product Hunt is a one-day spike,
and X is literally just other founders.
What's your plan to actually break out of the bubble and get real paying users?
English
@niccruzpatane That looks really small. I thought trucks in America were much longer / larger?
English
Bill Gates told Elon Musk that these were impossible.
Can’t wait to see next-gen Tesla Semis everywhere. 😎
Zanegler@HinrichsZane
Tesla Semi - Hatless Atlas See the refreshed Tesla semi doing calisthenics.
English
@asaio87 Because building scratches the creative itch. Selling requires facing rejection, talking to users, and doing things that don't scale. Most people choose the dopamine of shipping over the discomfort of selling. The 1% who do both are the ones who win.
English
I see a lot of people building with Claude Code or AI agents like crazy, but only 1% show their apps and have sales.
WHY ?
English
@tomdale My heart goes out to the engineer right now whose PM is insisting they “just ship” the MVP they vibecoded over the weekend
English
Down to speed and token cost really.
You could get it to do any number of steps on the code after the fact, but it all costs money to the end user.
I get they could just produce simpler code from the initial response, but that misses the point. You may be iterating on a solution and once you have that final solution you want to then clean it up
English
I see things like /simplify and the existence of code review and bug finding AIs.
I have to ask, why do these things exist? Why doesn't the coding agent just naturally do these things? I'm sure there's a good answer. Can someone help me understand?
Boris Cherny@bcherny
In the next version of Claude Code.. We're introducing two new Skills: /simplify and /batch. I have been using both daily, and am excited to share them with everyone. Combined, these kills automate much of the work it used to take to (1) shepherd a pull request to production and (2) perform straightforward, parallelizable code migrations.
English
I think the next killer feature from Vercel needs to be Auth + Billing for our Apps (like Clerk)
* Sign up users, all the components and integrations needed, allow Social signins, allow Vercel to manage the DB for it or allow me a provider so I can do it myself like Better Auth does
* Organizations + Member control
* Billing / Subscriptions with Org pricing, Per Seat Pricing and Usage tracking.
* For pricing side, allow me to setup a `product` key and allow me to sell it on my site and you handle all the heavy lifting, optionally allow me to process payments through any of the top providers (like Stripe)
So many apps would benefit from a Vercel provided framework and drop in shadcn style components like this
English
What service should we build next, with deep care and investment into its security, availability, and durability?
Guillermo Rauch@rauchg
Queues are one of the most requested services since I started Vercel. They're now here. It's just two APIs: 𝚜𝚎𝚗𝚍 and 𝚑𝚊𝚗𝚍𝚕𝚎𝙲𝚊𝚕𝚕𝚋𝚊𝚌𝚔 😌. The use-cases are basically infinite. Notably: queues can make agents and AI apps reliable. Quality and reliability are top of mind for everyone now, including our own team. We went through 3 iterations of the infrastructure while in private beta, and we're excited for you to build 'unbreakable software' with it.
English
@adamwathan Heya Adam, when does ui.sh launch, any hints? I am on the waitlist if you're accepting any folks early 🥰
English
So much of what makes a design looks visually great is matching your content to a design concept that fits your copy.
With ui.sh installed in your project, we'll point out to you when there's a mismatch and help you tweak the copy or find a better UI idea ✨
English
Gemini just one shotted this design
The original prompt is from @heyrobinai but I wanted to test it to see what he did and what level of detail was in the prompt.
English
@matteopelleg Not really fussed about the Apple side of this, but any model we can get running on consumer hardware that enables Agents to run will be the real winner
English
I’m now convinced that the biggest winner of the AI race will be Apple.
They will acquire Anthropic and put an AI model that can run on ~32GB of RAM in every device.
It will be private, local, have perfect memory, access to all of your files and it will cost $0.
The real moat was owning the hardware.
English
@chasedownleads Playing chess whilst everyone else is playing checkers
English
I took 217 Waymo rides yesterday
Each ride cost $7
I urinated in the backseat of every car
Guess who owns the company that charges $250 to thoroughly clean each car
$243 in pure profit for each ride
English
@levi Surely this is because vinext does not implement or support all the things that NextJS are doing under the hood
English
Tried switching from next.js to vinext and my dev reload times went from avg 2-5s to sub 100ms. The triangle company has some explaining to do.
English
@theo @jullerino Awesome. Will T3 Code support something like auto routing different requests to different models based on you detecting how indepth you think the request is?
English
Yesterday, @jullerino told me he had the Claude Code integration in T3 Code mostly done. I was surprised.
“Well, I’m down to 30% remaining in my Codex usage, so I wanted to use my Claude Code sub for easier tasks”
English