Follow AI Builders — Not Influencers

DeepSeek：世界上最耿直的模型

RT Alex Cui Claude's watermark probably doesn't work how you think. As the CTO of GPTZero, I'll explain how Anthropic, Google and OpenAI are building text watermarking in this brief explainer and whether it can be defeated. Almost all forms of watermarking that are fast and cheap enough for a frontier lab have the same formula, following the KGW method: In generation: 1. Let's say you've generated n tokens so far. Take those n tokens + a secret key to generate a random hash 2. Use that hash to randomly reweight the probabilities for the n+1 token, and then sample from that new distribution. In the simple case, you could split 50% of all English words into a green or red set based on your hash, and boost the probability of words in the green set. For watermark detection: 1. For each token, see if it was in the green or red set. 2. To do this, recreate the hash based on the secret key and the text preceding the current token. Then, recreate the green and red set of words. 3. Once you've checked all the words in the text, if the next token is selected disproportionally from the green set more than 50% of the time, you claim the text has the watermark. I can tell you want to ask the following: 1) Isn't it easy to mess up the hash if you paraphrase the text? The answer is mostly yes, however, you can use a statistical model to get your hash instead of a deterministic function (SIR, Adaptive Watermark). Since the entire watermark is probabilistic, this is fine. 2) Doesn't this make the text much worse? The answer is yes, it does - Yes, it does – but for most people, it's imperceptible (Google claims in human feedback study with 20,000 texts), since there are exponentially many ways to write the same paragraph. DiPmark does something more sophisticated to avoid shifting the text distribution on average. Of course, watermarks fail on short text or highly predictable texts like "2+2=4". 3) Shouldn't it be easy to figure out the green and red sets? The answer is no. ...

gpt luna max vs claude fable ultracode sent "pls build a mostly faithful clone of grok imagine with open models via fal" i woke up to these two and assumed fable was left and luna was right i was wrong... it was the other way!! objectively, fable did the better visual clone. but luna somehow understood intent better and created the more USABLE clone given my open model bent.

one man’s blur is another man’s motion data to decode

RT lily zhang sign up for COLM 2026 before it sells out! for non-local researchers, here's the SF map by @swyx. safety tip: i added poop alerts 💩 for the rough blocks. try to avoid booking hotel around union square and tenderloin, you'll thank me later. @COLM_conf

oh this is clean