frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Tested 31 AI detection/humanization tools – $5/mo GPTs beat $300/mo

1•khadinakbar•3w ago
I ran a systematic comparison of AI content detection and humanization tools after a client terminated a contract over an AI detection flag (87% AI-generated on content I'd manually edited).

*Methodology:* - 31 tools tested over 90 days - 200+ content samples (technical docs, marketing copy, blog posts, academic-style) - Measured detection accuracy against known AI/human content - Measured humanization "bypass rate" against Originality.ai (industry standard) - Controlled for content type and length

*Key finding:* ChatGPT Custom GPTs ($5/mo via team plans) performed within 2-7% of standalone SaaS tools charging $50-300/mo.

*Detection tools tested:* - Originality.ai: 91.3% accuracy, $149/mo unlimited - GPTZero: 87.4% accuracy, $16/mo - Copyleaks: 88.2% accuracy, $9-499/mo - Winston AI: 84.1% accuracy, $19/mo

*Humanization bypass rates (against Originality.ai):*

SaaS: - Undetectable.ai: 91.2%, $49-209/mo

Custom GPTs ($5/mo): - StealthGPT AI: 89.3% — https://chatgpt.com/g/g-67c88e5737388191aea00acc2e248afd - TurnitinPRO: 88.1% — https://chatgpt.com/g/g-67a36b4314548191a132428520afbf2d - BypassGPT: 87.6% — https://chatgpt.com/g/g-677e3f6ff8648191a96356838c564012 - ZeroGPT: 86.4% — https://chatgpt.com/g/g-67c88362d8e081918b73f42d780e53cb - GPT Zero: 86.2% — https://chatgpt.com/g/g-6786439fa24c81919660e0152ad5f4f3 - scribbr AI: 85.7% — https://chatgpt.com/g/g-67c89bebe2e48191962eaefb1e46530a - Humanize AI: 85.4% — https://chatgpt.com/g/g-674192227ff481918ff66a8dfe5378d9 - HumanizerPRO: 84.9% — https://chatgpt.com/g/g-67bfc9f5ab848191b7a80e386e7963af - Humanize AI Text: 84.7% — https://chatgpt.com/g/g-678cc08f1b048191a9428748d02916b1

*Cost comparison:*

Old stack: $223/mo - Originality.ai unlimited: $149 - Undetectable.ai: $49 - Quillbot: $10 - Grammarly: $15

New stack: $20/mo - ChatGPT Plus (team): $5 - Originality.ai pay-per-scan: ~$15

*Technical observations:*

1. Custom GPTs use the same base models as SaaS competitors. The differentiation is prompt engineering and workflow design, not proprietary detection/bypass algorithms.

2. Most humanizers fail on long-form content (>1500 words). Output becomes repetitive, tone drifts. BypassGPT and StealthGPT maintained consistency at 4000+ words.

3. Detection tools have different strengths: Originality.ai best overall accuracy, Copyleaks best for non-English content, GPTZero has more false positives on technical writing.

4. The "bypass rate" gap between $5 and $50+ tools (2-7%) matters less than workflow efficiency. Integrated detection+humanization in one interface saves ~30 min/article.

5. All tools struggle with heavily templated content (listicles, how-to formats). Detection accuracy drops 15-20% on these patterns regardless of actual AI involvement.

*Limitations:*

- Single tester, potential bias - Originality.ai as primary benchmark (other detectors may vary) - Custom GPT performance depends on OpenAI model updates - 90-day window; detection/bypass landscape evolves quickly

*Questions I'm still exploring:*

- How do detection tools handle fine-tuned models vs base GPT-4/Claude? - Is there a content length threshold where detection becomes unreliable? - How much does writing style (technical vs conversational) affect detection accuracy?

Comments

reify•3w ago
a fucking bargain

lots of meaningless figures

£300 quid a month.

Take a trip to Japan's Dododo Land, the most irritating place on Earth

https://soranews24.com/2026/02/07/take-a-trip-to-japans-dododo-land-the-most-irritating-place-on-...
1•zdw•26s ago•0 comments

British drivers over 70 to face eye tests every three years

https://www.bbc.com/news/articles/c205nxy0p31o
1•bookofjoe•47s ago•1 comments

BookTalk: A Reading Companion That Captures Your Voice

https://github.com/bramses/BookTalk
1•_bramses•1m ago•0 comments

Is AI "good" yet? – tracking HN's sentiment on AI coding

https://www.is-ai-good-yet.com/#home
1•ilyaizen•2m ago•1 comments

Show HN: Amdb – Tree-sitter based memory for AI agents (Rust)

https://github.com/BETAER-08/amdb
1•try_betaer•3m ago•0 comments

OpenClaw Partners with VirusTotal for Skill Security

https://openclaw.ai/blog/virustotal-partnership
1•anhxuan•3m ago•0 comments

Show HN: Seedance 2.0 Release

https://seedancy2.com/
1•funnycoding•3m ago•0 comments

Leisure Suit Larry's Al Lowe on model trains, funny deaths and Disney

https://spillhistorie.no/2026/02/06/interview-with-sierra-veteran-al-lowe/
1•thelok•3m ago•0 comments

Towards Self-Driving Codebases

https://cursor.com/blog/self-driving-codebases
1•edwinarbus•4m ago•0 comments

VCF West: Whirlwind Software Restoration – Guy Fedorkow [video]

https://www.youtube.com/watch?v=YLoXodz1N9A
1•stmw•5m ago•1 comments

Show HN: COGext – A minimalist, open-source system monitor for Chrome (<550KB)

https://github.com/tchoa91/cog-ext
1•tchoa91•5m ago•1 comments

FOSDEM 26 – My Hallway Track Takeaways

https://sluongng.substack.com/p/fosdem-26-my-hallway-track-takeaways
1•birdculture•6m ago•0 comments

Show HN: Env-shelf – Open-source desktop app to manage .env files

https://env-shelf.vercel.app/
1•ivanglpz•10m ago•0 comments

Show HN: Almostnode – Run Node.js, Next.js, and Express in the Browser

https://almostnode.dev/
1•PetrBrzyBrzek•10m ago•0 comments

Dell support (and hardware) is so bad, I almost sued them

https://blog.joshattic.us/posts/2026-02-07-dell-support-lawsuit
1•radeeyate•11m ago•0 comments

Project Pterodactyl: Incremental Architecture

https://www.jonmsterling.com/01K7/
1•matt_d•11m ago•0 comments

Styling: Search-Text and Other Highlight-Y Pseudo-Elements

https://css-tricks.com/how-to-style-the-new-search-text-and-other-highlight-pseudo-elements/
1•blenderob•13m ago•0 comments

Crypto firm accidentally sends $40B in Bitcoin to users

https://finance.yahoo.com/news/crypto-firm-accidentally-sends-40-055054321.html
1•CommonGuy•13m ago•0 comments

Magnetic fields can change carbon diffusion in steel

https://www.sciencedaily.com/releases/2026/01/260125083427.htm
1•fanf2•14m ago•0 comments

Fantasy football that celebrates great games

https://www.silvestar.codes/articles/ultigamemate/
1•blenderob•14m ago•0 comments

Show HN: Animalese

https://animalese.barcoloudly.com/
1•noreplica•14m ago•0 comments

StrongDM's AI team build serious software without even looking at the code

https://simonwillison.net/2026/Feb/7/software-factory/
3•simonw•15m ago•0 comments

John Haugeland on the failure of micro-worlds

https://blog.plover.com/tech/gpt/micro-worlds.html
1•blenderob•15m ago•0 comments

Show HN: Velocity - Free/Cheaper Linear Clone but with MCP for agents

https://velocity.quest
2•kevinelliott•16m ago•2 comments

Corning Invented a New Fiber-Optic Cable for AI and Landed a $6B Meta Deal [video]

https://www.youtube.com/watch?v=Y3KLbc5DlRs
1•ksec•18m ago•0 comments

Show HN: XAPIs.dev – Twitter API Alternative at 90% Lower Cost

https://xapis.dev
2•nmfccodes•18m ago•1 comments

Near-Instantly Aborting the Worst Pain Imaginable with Psychedelics

https://psychotechnology.substack.com/p/near-instantly-aborting-the-worst
2•eatitraw•24m ago•0 comments

Show HN: Nginx-defender – realtime abuse blocking for Nginx

https://github.com/Anipaleja/nginx-defender
2•anipaleja•24m ago•0 comments

The Super Sharp Blade

https://netzhansa.com/the-super-sharp-blade/
1•robin_reala•26m ago•0 comments

Smart Homes Are Terrible

https://www.theatlantic.com/ideas/2026/02/smart-homes-technology/685867/
2•tusslewake•27m ago•0 comments