Reddit

Built an AI workspace to simplify my SEO workflow — looking for honest feedback

Over the past few months, I've been building a project to solve a problem I kept running into. My SEO workflow was scattered across too many tools: Keyword research in one place SERP analysis in another Content briefs somewhere else AI writing in ChatGPT Competitor research across multiple tabs It felt like I was spending more time switching tools than actually creating content. So I started building a single workspace that brings these tasks together. Right now it can help w...
GLM-5 has 744B parameters and scores worse on MMLU-Pro than a 9B model
Reddit

GLM-5 has 744B parameters and scores worse on MMLU-Pro than a 9B model

Tier lists make S-tier and D-tier feel like different categories of thing entirely, red box at the top, blue box at the bottom. Actually plotted named models by parameter count against MMLU-Pro score instead of trusting the tier labels, and the picture is a lot messier than "bigger tier = bigger gap." Qwen3.5-9B, a 9B model, scores 82.5% on MMLU-Pro. GLM-5, at 744B parameters — 82x the size — scores 70.4%. That's not a diminishing-returns curve, that's negative returns; the 9B model ...
Reddit

Sinking of R.M.S. Titanic modelled using Fable 5

I wanted to better understand what happened hydraulically as the Titanic sank, so I created this simulation using Fable 5. The link shows the ship filling with water, breaking apart, the bow and stern ends sinking, and then impacting on the seafloor. No idea how accurate it is, but it is visually impressive and surprisingly polished.
Reddit

Thoughts on this ?

I got tired of seeing fly tipping near where I live so I started building an AI system to detect it. Computer vision, YOLOv8, trail cameras. 95% vehicle detection on first model. Building toward automatic alerts and evidence packaging for council prosecution. I’m 14 and doing this from my bedroom in Manchester.
Reddit

This week in AI: GPT-5.6, Gemini 3.5 Flash, Claude Science, and a Qwen price war — inference cost is collapsing across every tier at once

Lot dropped this week and there's a pretty clear through-line, so figured I'd pull it together. Model releases: - OpenAI launched GPT-5.6 (Sol/Terra/Luna). The bit worth noting isn't the flagship — it's Terra, reportedly matching GPT-5.5 quality at ~2x cheaper, with Luna aimed at the low-cost end. - Google shipped Gemini 3.5 Flash (beats 3.1 Pro on several benchmarks), plus Nano Banana 2 Lite (images ~$0.034/1K-res) and Gemini Omni Flash (video ~$0.10/sec via API). - xAI made Grok 3 GA a...
HackerNews

AI Poisons Creativity [video]

Article URL: https://www.youtube.com/watch?v=WeCSzEtZcUw Comments URL: https://news.ycombinator.com/item?id=48783386 Points: 2 # Comments: 1
HackerNews

Radio Free AI: The voice of the resistance

Article URL: https://open.spotify.com/show/033HWq6pzjHeleKR8jfxLm Comments URL: https://news.ycombinator.com/item?id=48783422 Points: 1 # Comments: 1
HackerNews

Ask HN: Which AI model do you use for what?

I use Grok for scanning X since it has direct access to the platform. Gemini for fact-checking. Claude for coding. GPT for image generation. Comments URL: https://news.ycombinator.com/item?id=48783556 Points: 1 # Comments: 0
HackerNews

Show HN: An MCP server that gives your AI assistant write access to /etc./hosts

I was tired of how bad distraction blockers are so I built my own. LockIn lets AI assistants (Claude, ChatGPT, Cursor, etc.) block websites by editing the system hosts file directly via MCP tools (via a secure cloudflare bridge). Why? Because chrome extensions don't work. Blocking it natively makes it unbypassable, and having your agent gatekeep your focus is the ultimate cheat code. Since everyone is using AI agents nowadays, but most people start scrolling or Reddit while waiting for Claude C...
HackerNews

Jersey Mike's IPO illustrates how bad the AI hype is

Article URL: https://finance.yahoo.com/technology/ai/articles/jersey-mike-ipo-illustrates-bad-201159743.html Comments URL: https://news.ycombinator.com/item?id=48784339 Points: 5 # Comments: 1
HackerNews

AI models' values are different from most people's

Article URL: https://www.economist.com/briefing/2026/06/25/ai-models-values-are-very-different-from-most-peoples Comments URL: https://news.ycombinator.com/item?id=48784353 Points: 2 # Comments: 0
HackerNews

Could the next great novel be written by AI?

Article URL: https://www.theguardian.com/books/ng-interactive/2026/jul/04/future-of-fiction-next-great-novel-ai-language-chat-gpt Comments URL: https://news.ycombinator.com/item?id=48784360 Points: 1 # Comments: 3
HackerNews

Teaching AI to Run with the Turbines

Article URL: https://www.technologyreview.com/2026/07/02/1138433/teaching-ai-to-run-with-the-turbines/ Comments URL: https://news.ycombinator.com/item?id=48784479 Points: 1 # Comments: 0
HackerNews

Show HN: ProxyBoy. A Windows HTTP/HTTPS debugging proxy with an AI assistant

ProxyBoy is a man-in-the-middle (MITM) HTTP/HTTPS proxy that captures, inspects, and modifies network traffic — similar to Charles Proxy, Fiddler, or Proxyman. What makes it different is the embedded AI assistant powered by the GitHub Copilot SDK, which can analyze traffic, create rules, and help debug network issues conversationally. Comments URL: https://news.ycombinator.com/item?id=48784551 Points: 1 # Comments: 0
Reddit

Do you agree with Palantir CEO Alex Karp that the enterprise "tokenmaxxing" business model has "gone completely wrong" with minimal ROI? Will open-weight models inevitably win?

Palantir CEO Alex Karp recently went on CNBC’s Squawk Box and delivered a brutal takedown of the API token pricing model pushed by commercial frontier labs like OpenAI and Anthropic. His core argument is that American enterprises are quietly "livid" because they are burning massive cash on skyrocketed token costs without seeing a clear return on investment. He noted that the industry’s incentive structure has completely devolved into meaningless "tokenmaxxing" —essential...
Reddit

Anthropic vs Opensourced model

Anthropic vs Open weight Chinese AI [ https://youtube.com/shorts/XZCWFNNiKgY?si=DViuG1xVptLTYDdQ\](https://youtube.com/shorts/XZCWFNNiKgY?si=DViuG1xVptLTYDdQ) When Alex Karp goes off on one of his rants, you usually have to filter through a lot of Palantir theater, but his recent take on AI safety was actually incredibly precise. He basically spelled out what real AI safety looks like for actual businesses, and it has nothing to do with vague alignment research or government certification ...