WANTED
Jacinto
Se busca por:
- Copy & Paste
- Tab to autocomplete AI
- Vibe Coding everything
- VLog in Youtube: One day as a software E.
- PR with conflicts
Se busca por:
<p><a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Farticles%2Fc4khcb83p8s23seu2u12.png" class="article-body-image-wrapper"><img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Farticles%2Fc4khcb83p8s23seu2u12.png" alt="Dashboard Preview" width="800" height="564"></a></p> <h2> canonical_url: <a href="https://github.com/guimaster97/pii-sanitizer-gateway" rel="noopener noreferrer">https://github.com/guimaster97/pii-sanitizer-gateway</a> </h2> <p>If you are building LLM wrappers or internal tools, you probably noticed two things killing your margins (and your sleep):</p> <ol> <li> <strong>Redundant API Costs:</strong> Users asking the same questions over and over, forcing you to pay OpenAI for the same tokens.</li> <li> <strong>Compliance Anxiety:</strong> The fear of a user accidentally pasting a client's Name, Email, or Tax ID into your chatbot, which then gets sent to a third-party server (OpenAI/DeepSeek).</li> </ol> <p>I looked for solutions, but most were heavy Enterprise Gateways (Java/Docker) or expensive SaaS. So, I decided to engineer my own solution running entirely on the Edge using <strong>Cloudflare Workers</strong>.</p> <p>Here is how I built <strong>Sanitiza.AI</strong>, an open-source gateway that caches requests and scrubs PII before they leave your network.</p> <h2> 🏗️ The Architecture </h2> <p>The goal was <strong>Zero DevOps</strong>. No Docker containers to manage, no Redis instances to pay for. Just pure Serverless functions.</p> <ul> <li> <strong>Runtime:</strong> Cloudflare Workers (TypeScript)</li> <li> <strong>Framework:</strong> Hono (A lightweight web framework, like Express but for the Edge)</li> <li> <strong>Storage:</strong> Cloudflare KV (Key-Value store for caching)</li> <li> <strong>Logic:</strong> Native Web Crypto API for hashing.</li> </ul> <h3> 1. The Money Saver: Smart Caching </h3> <p>For RAG (Retrieval-Augmented Generation) apps, redundancy is huge. To solve this, I implemented a "Smart Cache" mechanism.</p> <p>Before forwarding a request to OpenAI, the Worker creates a <strong>SHA-256 hash</strong> of the request body (the prompt + system instructions).</p> <div class="highlight js-code-highlight"> <pre class="highlight plaintext"><code> typescript // How we create a unique fingerprint for the request async function generateHash(message: string) { const msgBuffer = new TextEncoder().encode(message); const hashBuffer = await crypto.subtle.digest('SHA-256', msgBuffer); const hashArray = Array.from(new Uint8Array(hashBuffer)); return hashArray.map(b =&gt; b.toString(16).padStart(2, '0')).join(''); } It then checks Cloudflare KV. Hit: Returns the stored JSON instantly (&lt;50ms latency). Cost: $0. Miss: Forwards to OpenAI, gets the response, and saves it to KV for 24 hours. The Result: In my internal tools, I achieved a ~30% cache hit rate, effectively cutting my OpenAI bill by a third. 2. The Shield: PII Sanitization Sending PII (Personally Identifiable Information) to external LLMs is a GDPR/LGPD nightmare. I implemented a hybrid sanitization engine that runs before the request leaves the Edge. It uses Regex patterns (and can be extended with NER models) to identify sensitive entities. Example Input: "Analyze the contract for john.doe@email.com regarding the debt of $50k." What OpenAI Receives: "Analyze the contract for [EMAIL_HIDDEN] regarding the debt of [MONEY_HIDDEN]." The Worker maintains a mapping of these placeholders. When the LLM responds, the Worker "re-hydrates" the response, putting the real data back in if necessary, so the user never knows it was hidden. 3. The ROI Dashboard Engineering is useless if you can't prove value. I built a simple Admin Dashboard (served by the Worker itself via Hono) that tracks every token saved. (You can see the live calculator in the repo) 🚀 Performance on the Edge The biggest advantage of using Cloudflare Workers over a Python Proxy (like LiteLLM) is latency. The code runs in data centers close to the user. Cold Start: 0ms (practically). Cache Response Time: &lt; 50ms. Throughput: Cloudflare's network handles the load balancing. 💻 The Code (Open Source) I decided to open-source the core engine under the MIT license. You can deploy it to your own Cloudflare account in about 2 minutes using the "Deploy" button. It works with OpenAI, DeepSeek, Groq, and any other compatible API. Repository: github.com/guimaster97/pii-sanitizer-gateway I'm currently looking for contributors to help implement Semantic Caching (using Vectorize) to catch prompts that are similar but not identical. If you are into Rust/WASM or Vector Databases, let's talk! If you found this useful, drop a star on the repo. It helps a lot! </code></pre> </div>
<h2> Google Compares Gemini's Gradient Design to 1984 Macintosh </h2> <p>Recently, Google made an interesting comparison between the gradient design of Gemini, one of its latest products, and the iconic design of the Macintosh released in 1984. This comparison sparked a debate about the evolution of design in the tech world and how classic design elements can still influence modern products.</p> <h2> Background of the 1984 Macintosh </h2> <p>The 1984 Macintosh, officially known as the Macintosh 128k, was a revolutionary personal computer introduced by Apple. Released on January 24, 1984, this Macintosh was one of the first computers to use a graphical user interface (GUI) and gained widespread attention due to its iconic television commercial directed by Ridley Scott, which referenced George Orwell's dystopian novel, "1984". The commercial featured an athlete throwing a hammer at a screen displaying Big Brother's face, symbolizing that Macintosh would free users from IBM's dominance in personal computing.</p> <h2> Gemini's Gradient Design </h2> <p>Gemini, on the other hand, is Google's latest product featuring an attractive gradient design. While gradient design is not new in the world of design, its implementation on Gemini caught attention due to its resemblance to the 80s design aesthetic, including the 1984 Macintosh. The gradient on Gemini gives a dynamic and modern feel, showing how design can evolve while still retaining iconic elements from the past.</p> <h2> Google's Comparison </h2> <p>By comparing Gemini's gradient design to the 1984 Macintosh, Google highlights how design in technology continues to evolve in a circle. Although technology has advanced significantly over the past few decades, the principles of good design remain the same: creating intuitive, aesthetically pleasing, and functional interfaces. This comparison also shows how Google values design heritage in the tech industry and strives to integrate classic design elements into modern products.</p> <h2> Impact on the Design Industry </h2> <p>Google's comparison of Gemini's design to the 1984 Macintosh can have a significant impact on the design industry. It shows that designers don't always need to create something entirely new but can draw inspiration from iconic past designs and adapt them to meet modern needs and technology. This approach can help create products that are not only innovative but also have a classic touch that can be appreciated by users across different generations.</p> <h2> Conclusion </h2> <p>Google's comparison of Gemini's gradient design to the 1984 Macintosh highlights the evolution of design in the tech industry and how classic design elements can continue to influence modern products. It demonstrates that good design is about striking a balance between innovation and heritage, as well as the ability to draw inspiration from the past while looking towards the future.</p>
A new website takes Apple’s most persistent software bugs and reframes them as a hilarious scoreboard, using intentionally over-the-top math to estimate how many collective human hours they’ve supposedly wasted over the years. more…
Article URL: https://www.wsj.com/tech/ai/the-100-billion-megadeal-between-openai-and-nvidia-is-on-ice-aa3025e3 Comments URL: https://news.ycombinator.com/item?id=46831702 Points: 42 # Comments: 8
From time to time, OpenAI removes older models from ChatGPT’s model picker, as users become accustomed to newer ones. Here are the models being retired on February 13. more…
Apple is apparently pushing to turn its newly launched creative bundle into a brand of its own. Here are the details. more…
Don't lose your Fitbit data: New deadlines from Google you can't ignore.
A price so good it's hard to ignore
<p><a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Farticles%2Fc4khcb83p8s23seu2u12.png" class="article-body-image-wrapper"><img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Farticles%2Fc4khcb83p8s23seu2u12.png" alt="Dashboard Preview" width="800" height="564"></a></p> <h2> canonical_url: <a href="https://github.com/guimaster97/pii-sanitizer-gateway" rel="noopener noreferrer">https://github.com/guimaster97/pii-sanitizer-gateway</a> </h2> <p>If you are building LLM wrappers or internal tools, you probably noticed two things killing your margins (and your sleep):</p> <ol> <li> <strong>Redundant API Costs:</strong> Users asking the same questions over and over, forcing you to pay OpenAI for the same tokens.</li> <li> <strong>Compliance Anxiety:</strong> The fear of a user accidentally pasting a client's Name, Email, or Tax ID into your chatbot, which then gets sent to a third-party server (OpenAI/DeepSeek).</li> </ol> <p>I looked for solutions, but most were heavy Enterprise Gateways (Java/Docker) or expensive SaaS. So, I decided to engineer my own solution running entirely on the Edge using <strong>Cloudflare Workers</strong>.</p> <p>Here is how I built <strong>Sanitiza.AI</strong>, an open-source gateway that caches requests and scrubs PII before they leave your network.</p> <h2> 🏗️ The Architecture </h2> <p>The goal was <strong>Zero DevOps</strong>. No Docker containers to manage, no Redis instances to pay for. Just pure Serverless functions.</p> <ul> <li> <strong>Runtime:</strong> Cloudflare Workers (TypeScript)</li> <li> <strong>Framework:</strong> Hono (A lightweight web framework, like Express but for the Edge)</li> <li> <strong>Storage:</strong> Cloudflare KV (Key-Value store for caching)</li> <li> <strong>Logic:</strong> Native Web Crypto API for hashing.</li> </ul> <h3> 1. The Money Saver: Smart Caching </h3> <p>For RAG (Retrieval-Augmented Generation) apps, redundancy is huge. To solve this, I implemented a "Smart Cache" mechanism.</p> <p>Before forwarding a request to OpenAI, the Worker creates a <strong>SHA-256 hash</strong> of the request body (the prompt + system instructions).</p> <div class="highlight js-code-highlight"> <pre class="highlight plaintext"><code> typescript // How we create a unique fingerprint for the request async function generateHash(message: string) { const msgBuffer = new TextEncoder().encode(message); const hashBuffer = await crypto.subtle.digest('SHA-256', msgBuffer); const hashArray = Array.from(new Uint8Array(hashBuffer)); return hashArray.map(b =&gt; b.toString(16).padStart(2, '0')).join(''); } It then checks Cloudflare KV. Hit: Returns the stored JSON instantly (&lt;50ms latency). Cost: $0. Miss: Forwards to OpenAI, gets the response, and saves it to KV for 24 hours. The Result: In my internal tools, I achieved a ~30% cache hit rate, effectively cutting my OpenAI bill by a third. 2. The Shield: PII Sanitization Sending PII (Personally Identifiable Information) to external LLMs is a GDPR/LGPD nightmare. I implemented a hybrid sanitization engine that runs before the request leaves the Edge. It uses Regex patterns (and can be extended with NER models) to identify sensitive entities. Example Input: "Analyze the contract for john.doe@email.com regarding the debt of $50k." What OpenAI Receives: "Analyze the contract for [EMAIL_HIDDEN] regarding the debt of [MONEY_HIDDEN]." The Worker maintains a mapping of these placeholders. When the LLM responds, the Worker "re-hydrates" the response, putting the real data back in if necessary, so the user never knows it was hidden. 3. The ROI Dashboard Engineering is useless if you can't prove value. I built a simple Admin Dashboard (served by the Worker itself via Hono) that tracks every token saved. (You can see the live calculator in the repo) 🚀 Performance on the Edge The biggest advantage of using Cloudflare Workers over a Python Proxy (like LiteLLM) is latency. The code runs in data centers close to the user. Cold Start: 0ms (practically). Cache Response Time: &lt; 50ms. Throughput: Cloudflare's network handles the load balancing. 💻 The Code (Open Source) I decided to open-source the core engine under the MIT license. You can deploy it to your own Cloudflare account in about 2 minutes using the "Deploy" button. It works with OpenAI, DeepSeek, Groq, and any other compatible API. Repository: github.com/guimaster97/pii-sanitizer-gateway I'm currently looking for contributors to help implement Semantic Caching (using Vectorize) to catch prompts that are similar but not identical. If you are into Rust/WASM or Vector Databases, let's talk! If you found this useful, drop a star on the repo. It helps a lot! </code></pre> </div>
<h2> Google Compares Gemini's Gradient Design to 1984 Macintosh </h2> <p>Recently, Google made an interesting comparison between the gradient design of Gemini, one of its latest products, and the iconic design of the Macintosh released in 1984. This comparison sparked a debate about the evolution of design in the tech world and how classic design elements can still influence modern products.</p> <h2> Background of the 1984 Macintosh </h2> <p>The 1984 Macintosh, officially known as the Macintosh 128k, was a revolutionary personal computer introduced by Apple. Released on January 24, 1984, this Macintosh was one of the first computers to use a graphical user interface (GUI) and gained widespread attention due to its iconic television commercial directed by Ridley Scott, which referenced George Orwell's dystopian novel, "1984". The commercial featured an athlete throwing a hammer at a screen displaying Big Brother's face, symbolizing that Macintosh would free users from IBM's dominance in personal computing.</p> <h2> Gemini's Gradient Design </h2> <p>Gemini, on the other hand, is Google's latest product featuring an attractive gradient design. While gradient design is not new in the world of design, its implementation on Gemini caught attention due to its resemblance to the 80s design aesthetic, including the 1984 Macintosh. The gradient on Gemini gives a dynamic and modern feel, showing how design can evolve while still retaining iconic elements from the past.</p> <h2> Google's Comparison </h2> <p>By comparing Gemini's gradient design to the 1984 Macintosh, Google highlights how design in technology continues to evolve in a circle. Although technology has advanced significantly over the past few decades, the principles of good design remain the same: creating intuitive, aesthetically pleasing, and functional interfaces. This comparison also shows how Google values design heritage in the tech industry and strives to integrate classic design elements into modern products.</p> <h2> Impact on the Design Industry </h2> <p>Google's comparison of Gemini's design to the 1984 Macintosh can have a significant impact on the design industry. It shows that designers don't always need to create something entirely new but can draw inspiration from iconic past designs and adapt them to meet modern needs and technology. This approach can help create products that are not only innovative but also have a classic touch that can be appreciated by users across different generations.</p> <h2> Conclusion </h2> <p>Google's comparison of Gemini's gradient design to the 1984 Macintosh highlights the evolution of design in the tech industry and how classic design elements can continue to influence modern products. It demonstrates that good design is about striking a balance between innovation and heritage, as well as the ability to draw inspiration from the past while looking towards the future.</p>
Article URL: https://www.wsj.com/tech/ai/the-100-billion-megadeal-between-openai-and-nvidia-is-on-ice-aa3025e3 Comments URL: https://news.ycombinator.com/item?id=46831702 Points: 42 # Comments: 8
From time to time, OpenAI removes older models from ChatGPT’s model picker, as users become accustomed to newer ones. Here are the models being retired on February 13. more…
Apple is apparently pushing to turn its newly launched creative bundle into a brand of its own. Here are the details. more…
Just one day after the announcement of Genie, Google's generative AI-powered virtual world creator, a number of major video game companies are seeing their stock prices tumble, seemingly because investors think you can just generate an entire video game with AI now.
There was a genericness to the first wave of AI features in Google products. Absolutely everyone was doing straightforward text and image generation, and that contributed to the tacked-on nature. Fortunately, we’ve moved past that and Gemini integrations are starting to feel highly unique and play to Google’s strengths. more…
The Electronic Frontier Foundation (EFF) is out with a new campaign that presses tech companies to move faster to protect user data through end-to-end encryption, and stronger defaults and privacy settings. Here are the details. more…
A new website takes Apple’s most persistent software bugs and reframes them as a hilarious scoreboard, using intentionally over-the-top math to estimate how many collective human hours they’ve supposedly wasted over the years. more…
A price so good it's hard to ignore
Back in 2021, Nest Hub picked up an update that made it easy to access select functions via touch with a selection of “apps,” but they’ve vanished over the past year or so. more…
I love the smell of candles, but I hate the anxiety that comes with an open flame. There have been multiple times I’ve had to turn around after leaving to double-check that I’ve blown out a candle. That is why the SwitchBot Candle Warmer Lamp caught my eye. It promises all the ambiance and smells of a candle with zero risk of burning the house down. Best of all, it supports Matter right out of the box, making it a native citizen in your HomeKit/Apple Home environment. more…
Applicable this week onwards
Google Play Points has some new rewards music fans will love.
Parts of the app you don't even use look nice now.
Bloomberg’s Mark Gurman is well known for having numerous sources inside of Apple, and in a new interview he shared some interesting comments about the company’s reliance on Anthropic. more…
Skip the subscription
My zero-friction OneNote setup for maximum productivity
Google TV is testing out new “Free” and “Shop” tabs for the homescreen, with both tabs showcasing additional content in a slightly new way. more…
Oura won't be dropping its subscription model anytime soon.
Some Valley CEOs have condemned recent border patrol events, but Hoffman urges them to wield their influence more powerfully.
CSS Wrapped 2025 is out! We’re entering a world where CSS can increasingly handle logic, state, and complex interactions once reserved for JavaScript. Here is an unpacking of the standout highlights and how they connect to the bigger evolution of modern CSS.
Palmer Luckey thinks Meta closing its VR game studios is "a good thing for the long-term health of the industry", and that the narrative of it "abandoning" VR is "obviously false".
Team Beef, known for its standalone VR ports of classic games, is working on a VR version of Counter-Strike 1.6 for Quest and Pico headsets. The post An unofficial VR port of Counter-Strike 1.6 is coming to Meta Quest appeared first on MIXED Reality News.
A unique VR game that combines art and puzzles will be released in a few days - we are giving away free keys for this special app. The post Meta Quest: With a little luck, you can experience the striking VR puzzle Connectome for free appeared first on MIXED Reality News.
The Radeon 9070 XT is comparable to the RTX 5070 Ti.
The home theater experience is incomplete without great sound.
Plus, save on Valentine’s Day LEGO sets, score a bunch of MTG deals and preorders, and more.
Alienware's most powerful prebuilt gaming PC.
My RTX GPU is now punching up.
Donald Trump has appeared to undermine Second Amendment rights in statements about Alex Pretti’s killing. Many in the firearms community are going along with it.
https://github.com/omodaka9375/peerweb Comments URL: https://news.ycombinator.com/item?id=46829582 Points: 143 # Comments: 51
Lynx-R2, coming "this summer", is set to have the widest field of view of any standalone headset to date.
Don't lose your Fitbit data: New deadlines from Google you can't ignore.
The hacker allegedly developed zero-day exploits and offensive cyber tools and sold them to several countries, including an unnamed central African government, the U.K., and the United States.
The Polish government accused a Russian government hacking group of hacking into energy facilities taking advantage of default usernames and passwords.
Apple has discontinued the iPhone Upgrade Program in the UK, according to changes made to its website. The company says it is starting a “new chapter for upgrades” and that “the iPhone Upgrade Program is coming to an end.” Apple also appears to be winding down its partnership with Barclays, which was the bank behind the iPhone Upgrade Program in the UK. more…
It's 2026. What version of Android are YOU running?
Apple’s Wallet app received a variety of great features in iOS 26, but one of my favorites is a convenient new way to manage and access credit cards. Here’s how to use it. more…
After the Google Pixel got its monthly update in January, some users reported surging issues with both Bluetooth and Wi-Fi connectivity. Are you one of them? more…
Apple has released its annual Shot on iPhone short film in celebration of Chinese New Year. This year’s film, “Glad I Met You,” tells a heartwarming story of finding family with a stray dog. more…
Google’s mostly-forgotten Pixel Tablet is getting a new lease on life, as the company is giving its one and only tablet another two years of Android OS updates to match the policy on its phones. more…
Meta CTO Andrew Bosworth offered the first bit of insight into the company’s recent Reality Labs shakeup, publicly acknowledging that Meta’s metaverse efforts suffered from a “lack of focus” that ultimately hurt the user experience on Quest. Speaking at Axios House in Davos, Switzerland alongside the World Economic Forum last week, Bosworth discussed several issues […] The post Meta CTO: Metaverse Efforts Led to a “lack of focus” on Quest “at expense of user experience” appeared first on Road to VR.
Turn your favorite movies into magical moments.
Bigme’s B10 blends color E Ink, Android 14, and keyboard support.
A line in leaked log files suggests the upcoming flagship could could finally support Android’s Linux Terminal feature.
Since the feature's launch in 2018, users haven't been able to remove themselves from someone else's Close Friends list.
Today is the last day to lock in a plus-one pass at 50% off to TechCrunch Disrupt 2026, happening on October 13-15 in San Francisco.
With impeccable accuracy and poise, Samsung’s midrange OLED dazzles on the little things.
Alternate browser engines come to the Japanese App Store, but what will it take to see this in use? 🤔
Chrome DevTools MCP server, individual request throttling, adopted stylesheets and 2025 highlights.
Next.js 16.1 introduces Turbopack file system caching for development, a new bundle analyzer, and easier debugging in Node.js.
Discover some of the interesting features that have landed in stable and beta web browsers during December 2025.
Have you ever wished you could just poke the browser vendors and say, "Hey, I need this!"? Well, now you can.
Discover some of the interesting features that have landed in stable and beta web browsers during November 2025.