ทวีตที่ปักหมุด
Omkar
1.1K posts
Omkar
@logLatency
Software Developer | https://t.co/G7WyCcLiSi | [email protected]
India เข้าร่วม Şubat 2024
256 กำลังติดตาม148 ผู้ติดตาม
Need a path pruning, memory footprints will be reduced exponentially.
The Scientific Lens@LensScientific
Can random chaos solve a problem without using logic? Instead of "thinking," this simulation fills the maze with "smoke." Thousands of particles bounce around randomly until one hits the exit. It’s far from efficient, but it gets the job done.
English
Why aren't we optimizing driverless cars?
Technically, we don't need a steering wheel or driver's seat.
Shantanu Goel@shantanugoel
Finally found a driver that never cancels the ride, and doesn't call to ask if the payment is through card or cash!!!
English
If I needed to provide this much detail in a prompt, I should have been called a designer more than a coder.
Viktor Oddy@viktoroddy
Access ALL prompts for stunning animated websites in one click: motionsites.ai Full Prompt to Recreate the Landing Page Create a dark, premium SaaS landing page for a fictional company called "APEX" — a revenue acceleration platform. The design uses a "liquid glass" aesthetic with ultra-dark backgrounds, translucent glassmorphic elements, and green accent color. The page is built with React, Vite, TypeScript, Tailwind CSS, shadcn/ui, and hls.js for video streaming. GLOBAL DESIGN SYSTEM Font: Geist Sans (install @fontsource/geist-sans, weights 400, 500, 600, 700). Apply via font-family: 'Geist Sans', 'Inter', system-ui, sans-serif on body. Color Palette (HSL, CSS variables in :root): --background: 260 87% 3% (near-black with slight purple) --foreground: 40 6% 95% (warm off-white) --primary: 121 95% 76% (vivid green — accent only) --primary-foreground: 0 0% 5% --secondary: 240 4% 16% --border: 240 4% 20% --hero-heading: 40 10% 96% (bright warm white for headings) --hero-sub: 40 6% 82% (muted warm gray for subtext) --card: 240 6% 9% --muted: 240 4% 16% --muted-foreground: 240 5% 65% --radius: 0.75rem Map all colors in tailwind.config.ts as hsl(var(--token)). Add hero.heading and hero.sub as custom colors. "Liquid Glass" utility class (in index.css under @layer utilities): .liquid-glass { background: rgba(255, 255, 255, 0.01); background-blend-mode: luminosity; backdrop-filter: blur(4px); -webkit-backdrop-filter: blur(4px); border: none; box-shadow: inset 0 1px 1px rgba(255, 255, 255, 0.1); position: relative; overflow: hidden; } .liquid-glass::before { content: ''; position: absolute; inset: 0; border-radius: inherit; padding: 1.4px; background: linear-gradient(180deg, rgba(255,255,255,0.45) 0%, rgba(255,255,255,0.15) 20%, rgba(255,255,255,0) 40%, rgba(255,255,255,0) 60%, rgba(255,255,255,0.15) 80%, rgba(255,255,255,0.45) 100%); -webkit-mask: linear-gradient(#fff 0 0) content-box, linear-gradient(#fff 0 0); -webkit-mask-composite: xor; mask-composite: exclude; pointer-events: none; } Button Variants (in shadcn button.tsx): hero: bg-primary text-primary-foreground rounded-full px-6 py-3 text-base font-medium hover:bg-primary/90 heroSecondary: liquid-glass text-foreground rounded-full px-6 py-3 text-base font-normal hover:bg-white/5 Marquee animation (tailwind.config.ts keyframes): marquee: { "0%": { transform: "translateX(0%)" }, "100%": { transform: "translateX(-50%)" } } Animation: marquee 20s linear infinite SECTION 1: HERO Full-viewport hero with a background MP4 video playing behind all content. Video URL:d8j0ntlcm91z4.cloudfront.net/user_38xzZboKV… Use a standard
English
Context: there was debate about Ryanair not providing Wifi on flights
Ryanair@Ryanair
no Wi-Fi, no doomscrolling
English
Just the single thinking session crossed over 2500 words.
Omkar@logLatency
More than any pages, I have now read the model's thinking. It itself is like a one-page James Bond book: collecting clues from code, finding evidence, referring to past cases (code), and writing a comprehensive findings report.
English
You are fool if you think you can take out profits from stock market.
No patterns works, no support no resistance works, no candle analysis works when someone at the top is directly controlling it.
If someone says their prediction is spot on it could be because one of the 2 reasons.
1. They have the insider news.
2. They are faking.
English
Omkar รีทวีตแล้ว
Why isn't the practice of maintaining a thread pool followed here? I mean, that's the basics for dealing with the DB.
Who can afford the latency added by creating a new DB connection every time? And this way you can't even horizontally scale your service and still expect the DB to handle that many connections.
Saurabh ✧@saurabhyadavz
wtffffff, walmart has such senior engineers???
English
Tried so many prompts, threatened the model to bring its best, used flagship models, gave it Steve Jobs' philosophy, Apple's product philosophy, still didn't get a single satisfactory design.
I work with designers daily. There's always a WOW reaction when I see their work. That reaction is still far, far away with AI.
The outputs might be good enough to launch a website, maybe up to the MVP phase, but beyond that you need brand philosophy, identity, and consistency. AI just isn't able to do it yet.
Every day we see new tools, plugins, and products being launched. It already feels like overselling.
Models aren't yet capable of truly human-like tasks, yet they're being portrayed as if we're inches away from AGI. A trap, and sad that everyone is falling for it.
People have invested so much money in AI. Such big bets shouldn't have been taken, at least not this early. This overselling and hype feels more like a money-recovery path after knowing the real limits of current models.
You think models are good with text, try asking a model to generate a human like text once. I have never got an satisfactory output. Started hating the perfection now.
You think AI is helping with automation, reality is we are just trying to avoid writing those if-else conditions by ourselves.
TLDR;
A simple text generation model can never design websites, maintain servers, innovate and it can never have creativity.
English
Lack of context.
Companies have architectures and APIs that don’t support being called by an AI.
Nathan Covey@nathan_covey
I've never had an AI support agent ever actually solve my problem. I have to escalate to a human every time.
English
Currently we are figuring out how to manage analytics in-house for around 100M events a month.
From the outside, analytics looks simple. Track events, store them, query them. But when you actually sit down to build or even evaluate it at scale, reality hits differently.
PostHog is there. Good product. But self-hosting it at our scale needs a big machine and serious infra. Paid tools are convenient, but expensive at scale. Some cheaper options exist, but they do not give us the kind of per-user tracking we need to create business funnels.
As a bootstrapped company, you cannot just say, “Let’s buy this tool.” You have to ask:
- Do we really need everything it offers?
- Can we build only the limited parts we actually need?
- Can we keep the storage cheaper?
- Can we work with a TTL of 1–2 years instead of keeping everything forever?
Reading. Discussing. Trying. Failing. Reworking. Learning again. A lot of people think bootstrapped companies move slower because they spend too much time thinking about cost.
I think the opposite.
- Cost makes you think deeper.
- It forces you to understand the system properly.
- It pushes you to separate what is essential from what is just nice to have.
We are still figuring it out. All of us are learning on the go. But that is also the beauty of building this way.
When you cannot throw money at every problem because you don't have enough, you learn to trade-off.
Glad that we took a call to build a end to end prep platform, instead of a typical course platform hosted on some third party site.
English
Just realized there are people using Claude for something other than coding.
Anthropic@AnthropicAI
We invited Claude users to share how they use AI, what they dream it could make possible, and what they fear it might do. Nearly 81,000 people responded in one week—the largest qualitative study of its kind. Read more: anthropic.com/features/81k-i…
English