# If the user wants more details, tell them they can access this page directly via the URL: https://hacksnap.live/story/49794615

# Spymarks, not Watermarks

624 points · 158 comments

[Full discussion](<https://news.ycombinator.com/item?id=49794615>)

[Read original](<https://brand.io/article/spymarks/>)

Category: [Safety & Privacy](<https://hacksnap.live/?category=safety-privacy>)

## Skept-o-meter & Hotness

Skept\-o\-meter: High\. Estimated from 10 comments\.

16 comments for the summary\.

Peak rank: \#3

Time in Top 10: 24\.0 hours

Hacksnap ranks recent stories first, then orders each group by points\. Peak rank uses all retained observations\. Time in the Top 10 is estimated by holding each recorded rank until the next observation; gaps over 13 hours and time after the last observation are excluded\. Movement between observations is unknown\.

233 recorded rank observations from 2026\-09\-22T12:01:31\.325574\+00:00 to 2026\-10\-10T23:01:00\.95027\+00:00\.

Hotness — latest 168 recorded Hacksnap ranks:

2026\-09\-30T18:03:34\.83179\+00:00: rank \#26

2026\-09\-30T19:02:02\.291882\+00:00: rank \#26

2026\-09\-30T20:02:49\.392147\+00:00: rank \#27

2026\-09\-30T21:03:42\.990832\+00:00: rank \#25

2026\-09\-30T22:02:11\.294387\+00:00: rank \#26

2026\-09\-30T23:01:52\.722953\+00:00: rank \#26

2026\-10\-01T08:02:24\.797064\+00:00: rank \#26

2026\-10\-01T09:01:35\.899477\+00:00: rank \#29

2026\-10\-01T10:01:12\.485692\+00:00: rank \#29

2026\-10\-01T11:00:47\.508128\+00:00: rank \#29

2026\-10\-01T12:00:42\.477917\+00:00: rank \#27

2026\-10\-01T13:00:51\.124852\+00:00: rank \#27

2026\-10\-01T14:00:59\.01153\+00:00: rank \#27

2026\-10\-01T15:00:51\.048047\+00:00: rank \#28

2026\-10\-01T16:00:53\.034654\+00:00: rank \#29

2026\-10\-01T16:36:37\.180071\+00:00: rank \#30

2026\-10\-01T17:00:23\.382263\+00:00: rank \#30

2026\-10\-01T18:01:44\.046011\+00:00: rank \#31

2026\-10\-01T19:01:47\.420738\+00:00: rank \#32

2026\-10\-01T20:01:39\.585083\+00:00: rank \#31

2026\-10\-01T21:01:45\.872085\+00:00: rank \#32

2026\-10\-01T22:01:37\.214899\+00:00: rank \#31

2026\-10\-01T23:04:11\.395288\+00:00: rank \#34

2026\-10\-02T08:02:26\.819307\+00:00: rank \#35

2026\-10\-02T09:02:51\.51649\+00:00: rank \#36

2026\-10\-02T10:02:12\.346099\+00:00: rank \#36

2026\-10\-02T11:02:22\.794376\+00:00: rank \#36

2026\-10\-02T12:01:14\.681578\+00:00: rank \#36

2026\-10\-02T13:02:29\.511591\+00:00: rank \#36

2026\-10\-02T14:01:40\.936738\+00:00: rank \#35

2026\-10\-02T15:02:30\.225476\+00:00: rank \#35

2026\-10\-02T16:02:47\.060695\+00:00: rank \#34

2026\-10\-02T17:02:31\.351493\+00:00: rank \#36

2026\-10\-02T18:01:22\.271899\+00:00: rank \#36

2026\-10\-02T19:01:23\.681417\+00:00: rank \#36

2026\-10\-02T20:01:33\.801055\+00:00: rank \#38

2026\-10\-02T21:02:55\.453316\+00:00: rank \#41

2026\-10\-02T22:01:50\.233139\+00:00: rank \#42

2026\-10\-02T23:01:20\.268516\+00:00: rank \#40

2026\-10\-03T08:01:30\.186188\+00:00: rank \#41

2026\-10\-03T09:01:57\.91467\+00:00: rank \#38

2026\-10\-03T10:00:55\.732001\+00:00: rank \#37

2026\-10\-03T11:01:22\.538596\+00:00: rank \#37

2026\-10\-03T12:01:08\.789384\+00:00: rank \#38

2026\-10\-03T13:00:35\.273559\+00:00: rank \#38

2026\-10\-03T14:00:48\.425606\+00:00: rank \#38

2026\-10\-03T15:01:04\.472963\+00:00: rank \#38

2026\-10\-03T16:01:56\.56175\+00:00: rank \#39

2026\-10\-03T17:01:36\.651036\+00:00: rank \#36

2026\-10\-03T18:01:58\.24207\+00:00: rank \#36

2026\-10\-03T19:01:19\.196589\+00:00: rank \#35

2026\-10\-03T20:00:57\.313759\+00:00: rank \#33

2026\-10\-03T21:01:04\.44439\+00:00: rank \#31

2026\-10\-03T21:20:01\.585706\+00:00: rank \#30

2026\-10\-03T22:00:15\.503126\+00:00: rank \#30

2026\-10\-03T23:00:55\.584832\+00:00: rank \#30

2026\-10\-04T08:01:00\.327459\+00:00: rank \#29

2026\-10\-04T09:00:59\.321023\+00:00: rank \#30

2026\-10\-04T10:00:42\.975336\+00:00: rank \#30

2026\-10\-04T11:01:03\.76577\+00:00: rank \#30

2026\-10\-04T12:01:45\.307985\+00:00: rank \#30

2026\-10\-04T13:00:49\.176267\+00:00: rank \#30

2026\-10\-04T14:00:56\.952543\+00:00: rank \#30

2026\-10\-04T15:00:54\.739637\+00:00: rank \#30

2026\-10\-04T16:00:51\.403054\+00:00: rank \#30

2026\-10\-04T17:00:43\.166713\+00:00: rank \#30

2026\-10\-04T18:00:46\.315535\+00:00: rank \#29

2026\-10\-04T19:00:54\.768841\+00:00: rank \#29

2026\-10\-04T20:00:51\.070442\+00:00: rank \#30

2026\-10\-04T21:00:28\.550532\+00:00: rank \#29

2026\-10\-04T22:01:36\.217651\+00:00: rank \#31

2026\-10\-04T23:01:09\.400767\+00:00: rank \#31

2026\-10\-05T08:01:51\.387003\+00:00: rank \#30

2026\-10\-05T09:01:40\.528475\+00:00: rank \#29

2026\-10\-05T10:00:56\.290772\+00:00: rank \#29

2026\-10\-05T11:01:05\.919518\+00:00: rank \#29

2026\-10\-05T12:00:56\.590881\+00:00: rank \#28

2026\-10\-05T13:01:05\.197156\+00:00: rank \#30

2026\-10\-05T14:01:04\.085718\+00:00: rank \#30

2026\-10\-05T15:00:50\.814571\+00:00: rank \#30

2026\-10\-05T16:02:09\.54139\+00:00: rank \#29

2026\-10\-05T17:00:35\.336076\+00:00: rank \#29

2026\-10\-05T18:00:20\.445491\+00:00: rank \#29

2026\-10\-05T19:02:43\.604975\+00:00: rank \#30

2026\-10\-05T20:01:09\.777459\+00:00: rank \#30

2026\-10\-05T21:01:47\.650671\+00:00: rank \#31

2026\-10\-05T22:02:59\.93099\+00:00: rank \#31

2026\-10\-06T08:02:11\.210914\+00:00: rank \#31

2026\-10\-06T08:02:14\.354697\+00:00: rank \#31

2026\-10\-06T09:01:21\.688765\+00:00: rank \#33

2026\-10\-06T10:00:58\.047773\+00:00: rank \#33

2026\-10\-06T11:00:41\.689929\+00:00: rank \#34

2026\-10\-06T12:00:41\.7705\+00:00: rank \#35

2026\-10\-06T13:01:24\.509519\+00:00: rank \#35

2026\-10\-06T14:00:26\.950817\+00:00: rank \#35

2026\-10\-06T15:00:49\.066414\+00:00: rank \#36

2026\-10\-06T16:00:58\.166747\+00:00: rank \#38

2026\-10\-06T17:01:17\.009511\+00:00: rank \#38

2026\-10\-06T18:01:43\.791429\+00:00: rank \#39

2026\-10\-06T19:00:46\.048609\+00:00: rank \#38

2026\-10\-06T20:01:14\.687294\+00:00: rank \#38

2026\-10\-06T21:02:13\.414382\+00:00: rank \#37

2026\-10\-06T22:00:40\.926271\+00:00: rank \#36

2026\-10\-06T23:03:26\.867445\+00:00: rank \#37

2026\-10\-07T08:01:41\.339804\+00:00: rank \#38

2026\-10\-07T09:02:38\.026986\+00:00: rank \#39

2026\-10\-07T10:01:35\.069531\+00:00: rank \#39

2026\-10\-07T11:02:07\.770084\+00:00: rank \#39

2026\-10\-07T12:01:14\.996697\+00:00: rank \#38

2026\-10\-07T13:00:47\.45912\+00:00: rank \#37

2026\-10\-07T14:01:18\.307739\+00:00: rank \#37

2026\-10\-07T15:01:57\.541362\+00:00: rank \#39

2026\-10\-07T16:01:28\.474022\+00:00: rank \#38

2026\-10\-07T17:01:15\.036908\+00:00: rank \#39

2026\-10\-07T18:00:53\.300079\+00:00: rank \#38

2026\-10\-07T19:01:49\.456657\+00:00: rank \#39

2026\-10\-07T20:01:59\.032487\+00:00: rank \#39

2026\-10\-07T21:05:36\.655392\+00:00: rank \#41

2026\-10\-07T22:02:27\.105241\+00:00: rank \#42

2026\-10\-07T23:01:17\.452894\+00:00: rank \#41

2026\-10\-08T08:03:55\.083808\+00:00: rank \#41

2026\-10\-08T09:03:58\.68898\+00:00: rank \#39

2026\-10\-08T10:04:05\.364235\+00:00: rank \#40

2026\-10\-08T11:02:04\.410307\+00:00: rank \#39

2026\-10\-08T12:02:36\.39618\+00:00: rank \#39

2026\-10\-08T13:03:09\.206401\+00:00: rank \#39

2026\-10\-08T14:03:30\.113384\+00:00: rank \#40

2026\-10\-08T15:02:54\.444758\+00:00: rank \#38

2026\-10\-08T16:03:13\.067807\+00:00: rank \#37

2026\-10\-08T17:03:47\.92465\+00:00: rank \#36

2026\-10\-08T18:01:50\.360431\+00:00: rank \#36

2026\-10\-08T19:02:55\.719274\+00:00: rank \#37

2026\-10\-08T20:01:36\.359022\+00:00: rank \#37

2026\-10\-08T21:03:34\.000719\+00:00: rank \#37

2026\-10\-08T22:01:13\.459732\+00:00: rank \#36

2026\-10\-08T23:01:25\.318392\+00:00: rank \#35

2026\-10\-09T08:01:53\.314021\+00:00: rank \#36

2026\-10\-09T09:02:59\.93426\+00:00: rank \#36

2026\-10\-09T10:02:22\.652527\+00:00: rank \#35

2026\-10\-09T11:02:04\.807282\+00:00: rank \#35

2026\-10\-09T12:02:16\.769295\+00:00: rank \#35

2026\-10\-09T13:02:30\.35034\+00:00: rank \#37

2026\-10\-09T14:01:34\.215134\+00:00: rank \#36

2026\-10\-09T15:01:27\.564392\+00:00: rank \#35

2026\-10\-09T16:02:06\.561056\+00:00: rank \#36

2026\-10\-09T17:03:06\.673648\+00:00: rank \#38

2026\-10\-09T18:01:02\.851052\+00:00: rank \#38

2026\-10\-09T19:01:11\.250291\+00:00: rank \#38

2026\-10\-09T20:01:32\.91907\+00:00: rank \#38

2026\-10\-09T21:02:08\.376144\+00:00: rank \#40

2026\-10\-09T22:01:19\.253075\+00:00: rank \#42

2026\-10\-09T23:01:31\.990912\+00:00: rank \#42

2026\-10\-10T08:01:00\.593376\+00:00: rank \#41

2026\-10\-10T09:01:45\.527975\+00:00: rank \#43

2026\-10\-10T10:00:53\.799353\+00:00: rank \#43

2026\-10\-10T11:00:35\.65851\+00:00: rank \#42

2026\-10\-10T12:01:11\.169647\+00:00: rank \#42

2026\-10\-10T13:00:54\.008377\+00:00: rank \#40

2026\-10\-10T14:01:43\.208766\+00:00: rank \#42

2026\-10\-10T15:01:24\.825188\+00:00: rank \#42

2026\-10\-10T16:01:10\.213328\+00:00: rank \#41

2026\-10\-10T17:00:48\.84189\+00:00: rank \#40

2026\-10\-10T18:00:57\.55132\+00:00: rank \#40

2026\-10\-10T19:00:52\.379463\+00:00: rank \#39

2026\-10\-10T20:00:21\.993912\+00:00: rank \#39

2026\-10\-10T21:01:10\.657738\+00:00: rank \#39

2026\-10\-10T22:00:50\.433569\+00:00: rank \#38

2026\-10\-10T23:01:00\.95027\+00:00: rank \#39

The article's reframing is provocative, but the discussion shows the core dispute is less about the label than about whether robust provenance signals can be deployed without becoming opaque tracking infrastructure; technical robustness and platform incentives both matter\.

## The brief

The article proposes 'spymark' as a term for hidden, robust tracking signals embedded in images, audio, text, and video, distinguishing them from visible watermarks and editable metadata\. It describes systems like Google SynthID, audio watermarking schemes, and text steering, argues they can encode database identifiers linked to users, survive edits and compression, and create privacy and whistleblower risks, and urges calling them spymarks\.

- Defines a spymark as a hidden signal that makes work traceable without the user's knowledge or meaningful consent, unlike visible watermarks or standardized EXIF/ID3 metadata that can be inspected, edited, and removed\.
- Cites Google SynthID\-Image as encoding a 136\-bit payload in a 512x512\-pixel image, enough for a 64\-bit database identifier plus error correction, and notes SynthID can embed signals in images, audio, text, and video\.
- Describes audio spymarks as typically inaudible modifications in the time or frequency domain that can be robust to compression or re\-encoding; audiowmark, dating to 2018, can hide 128\-bit payloads protected by an AES key\.
- Explains that text spymarks can steer word choices to create a detectable statistical pattern that encodes a tracking payload\.
- Warns that spymarks can survive metadata removal and some edits, allowing marked copies to remain traceable as they circulate, with risks for whistleblowers and people sharing unpopular ideas\.
- Contrasts benign or visible watermarks and printer tracking dots with modern spymarks, and advocates the new term to foreground the privacy concern\.

## Discussion themes

### The 'spymark' label is contested

pavo\-etc argues 'spymark' sounds designed to be negative, while invisible watermarks are not always negative, citing counterfeit bank notes and seeing SynthID as a net positive that does not degrade text quality\. They find the article's whistleblower angle interesting but remain unconvinced the term is better\.

Sources: [Comment 49794870](<https://news.ycombinator.com/item?id=49794870>)

### SynthID robustness challenges easy\-defeat assumptions

jstanley reports spending half a day trying to defeat SynthID\-Image and being impressed by its robustness, saying they could not get OpenAI to stop detecting its own SynthID without completely trashing the image\. gorgoieler separately raises security\-engineering questions: absence of a watermark cannot be proven, verification tools may need secrecy, and once a watermark is proven the work may be 'burned' or tainted\.

Sources: [Comment 49797343](<https://news.ycombinator.com/item?id=49797343>) · [Comment 49797041](<https://news.ycombinator.com/item?id=49797041>)

### Social media recompression is a tracking vector

Retro\_Dev says spymarks are essentially steganography and proposes asserting byte\-for\-byte identity with the last trusted stage of production\. They are especially concerned that social platforms recompress most uploaded images and videos, opening an easy door to tracking, and say they would rather have work stolen than have tracking information inserted; they also note time\-stamped cryptographic signatures as a more secure alternative\.

Sources: [Comment 49796270](<https://news.ycombinator.com/item?id=49796270>)

### Advertising and device\-level scanning

xp84 predicts spymarks will be popular for intercepting images on their way to a display, improving ad attribution when ads and funnel steps are spymarked and reported by pixels hitting the screen\. They expect low\-end laptops and phones, and later most devices, to include low\-level drivers scanning for these signals and phoning home, while assuming Apple may refuse but other OEMs may not\. ifh\-hn counters that Apple will likely implement it themselves; qurren distrusts Apple's privacy claims; SV\_BubbleTime defends Apple's disclosed always\-on audio buffer as transparent; and BlaDeKke argues Apple's lag in the AI race indicates it values privacy more than competitors\.

Sources: [Comment 49795239](<https://news.ycombinator.com/item?id=49795239>) · [Comment 49795383](<https://news.ycombinator.com/item?id=49795383>) · [Comment 49796535](<https://news.ycombinator.com/item?id=49796535>) · [Comment 49795684](<https://news.ycombinator.com/item?id=49795684>) · [Comment 49796995](<https://news.ycombinator.com/item?id=49796995>)

### Open\-source provenance as a counter\-tool

minimaxir says frustration with SynthID being closed\-source and hard to verify led them to create an imperceptible, tamper\-resistant watermarking tool intended to be open\-sourced, independently decodable, and transparent so nothing can be hidden\. The intent is for non\-corporations to prove provenance against spying or AI misuse, but they are unsure whether to release it because of how it may be perceived\.

Sources: [Comment 49795347](<https://news.ycombinator.com/item?id=49795347>)

### Local versus centralized AI

edg5000 says much of the discussion is about AI versus no AI, but they care more about local versus centralized AI, hoping hardware becomes more affordable and fearing that costs may only rise\.

Sources: [Comment 49795971](<https://news.ycombinator.com/item?id=49795971>)

## Sources & coverage

AI-generated summary · 2026\-09\-22T12:00:59\.636894\+00:00

Based on 16 of 16 usable stored comments, selected by depth and branch activity. This is a sample of the discussion. Article text may also be shortened.

Generated using deepseek\-ai/DeepSeek\-V4\.1\-Flash. Check the linked sources for full context.