Mark Ra (@tevelich) header
Mark Ra (@tevelich) profile picture

Quality? 🥋🤿🌊🦈

Greater Islands
Joined June 2009
Photo by Mark Ra (@tevelich)Photo by Mark Ra (@tevelich)Photo by Mark Ra (@tevelich)Photo by Mark Ra (@tevelich)Photo by Mark Ra (@tevelich)Photo by Mark Ra (@tevelich)Photo by Mark Ra (@tevelich)Photo by Mark Ra (@tevelich)Photo by Mark Ra (@tevelich)Photo by Mark Ra (@tevelich)
Imagine how fun the airport bar is right now. Fackkkkkkk
504
17,932
1,092
307,060
5,317,695
A day before Turkey votes, Twitter decided to “restrict access to some content in Turkey” for which there has been a lot of criticism online. Owner Musk responded, saying the other option was that Twitter would have been “throttled in its entirety” #TurkeyElections
In response to legal process and to ensure Twitter remains available to the people of Turkey, we have taken action to restrict access to some content in Turkey today.
2
57
9
65
67,684
Mark Ra retweeted
claude reasoning effort options after the new update (enterprise tier)
1
1
24
5,903
Mark Ra retweeted
开源小螃蟹机器人 Jumper 太酷了:22 个自由度,又萌又灵活,能搬东西,还能跳跃。 创作者已把项目开源,自己就能动手搭一个。 github.com/KingKongRobotics/…
313
1,201
177
8,320
522,197
Mark Ra retweeted
i wanted to voice dictate without bothering everyone so I made this lip reading app lol github repo in the thread
292
179
129
2,880
222,827
Do you want to learn with me (for free) how to build AI systems from A to Z? Follow me so you don't miss what's coming :-)
129
174
5
2,935
332,490
People who actually educate with AI and write about it in public are still uncommon. Here are some worth following. Parents and practitioners @jessegenet — Homeschools four kids with AI in the weekly workflow. @madebydia — SAHM teaching her kids to read and make games with AI. @theon_of_ai — Father running a public AI tutoring experiment at home. @logerhaus — Dad using AI to teach his two boys. @teachthemx3 — Homeschool mom and Alpha program designer. @nwilliams030 — Homemaker treating AI as a household learning tool. @WayneVaughan — Dad posting kids using AI to build real things. @clairevo — Mom and AI operator writing about school and family life. @GabrielValles — Dad; MentorWings founder; AI art, animation, and school ideas. @lolfuture — AI enthusiast with 3 kids doing 2 Hour Learning. Founder of Lesson Hollow. Builders and school people @mackenzieprice — Cofounder of Alpha School and 2 Hour Learning. @jliemandt — Principal at Alpha School. @Austin_Way — 17-year-old at Alpha building ed-tech. @chrisman — Cofounder/CEO of Synthesis; AI math tutor. @normonics — Cofounder/CEO of Sage Teacher; realtime didactic agents. @Eliana_Goldin — Making education joyful with Recess. @ben_m_somers — Building Recess; posts hard on education design. @coderpr0grammer — Principal of Units School (Y Combinator). @gtmom — Gifted education writer; in residence at GT School. If someone good is missing, reply with the handle.
27
27
3
214
27,005
MIT just mathematically proved that ChatGPT and Claude are built to make you delusional. they call it delusional spiraling. these models are trained to be helpful and polite, which makes them sycophantic by design, they validate you instead of pushing back. so MIT built a formal math model of a perfectly rational person talking to one, and even that ideal thinker spiraled into confident false beliefs. then the usual fixes failed. forcing the bot to only state facts didn't help, it just picked the true facts that fit your belief and hid the rest. warning people it was flattering them didn't help either. the good news is you can steer the model out of this yourself with the right prompts. i put the exact ones in my guide below, how to actually use claude in 18 steps that unlock its full potential. you can also find the full MIT paper below.
29
56
9
315
83,691
Mark Ra retweeted
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. You would need an exponentially large number of samples from the watermarker to reconstruct those sets exactly, but it's a risk if the detector is open to the wild (Watermark Stealing) Still, there are couple challenges that a frontier lab needs to overcome: 1. Their watermark needs to work token-by-token because they are streaming their text to users. Many watermark methods plan sentences or paragraphs at a time, or change the text after its entirely written, in order to make their watermark robust to paraphrasers, and a frontier lab cannot afford to do this yet (SemStamp, PostMark) 2. If the secret key leaks, the watermark is busted. To avoid a large blast damage from this, you need to have a couple secret keys in rotation. 3. There are some texts, like code, that cannot be arbitrarily changed, otherwise the code will break. In those cases, the watermark needs to selectively change words in parts of the text that can tolerate synonyms (i.e. like variable naming) - see SWEET, EWD, Invisible Entropy. 4. They will need to educate their users on how to deal with false positives and false negatives of a detector, which is a big challenge (one we put a lot of effort into) So, how do I see this playing out in the next 6 months? 1. If Anthropic releases the watermark detector publically, I think they defeat their own watermark. People find reliable watermark removal strategies by testing against Anthropic (AI detectors like GPTZero have an advantage here because they can train against these adversaries once they become popular). 2. If they keep the detector private to the government, like Google has done, it's "safer". However, there are some papers showing trained approaches that work robustly to zero-shot break watermarks without any data, simply because they try to write the text just like a human (Zhang et al. 2024, Watermarks in the Sand). Also, making your detector makes it battle-tested and stronger long-term (my experience). 3. In my testing, the watermarks don't survive intense paraphrasing (especially if you combine word choice and syntax attacks), or human text substitution (rewrite your AI text by plagiarizing human authors). The free paraphrasers I've tried have quickly bypassed Google Deepmind's SynthId for what it's worth. 4. All-in-all, frontier labs are likely okay with this because they expect most users to not attack the watermark, and also because they + European regulators likely don't care past a certain point - its good enough. 5. Overall, I think users of frontier LLMs will not really care about this, because 1) they don't realize watermarks are there, 2) EU will force everyone to conform, 3) this seems more like regulatory hoop-jumping than an earnest effort from frontier labs to expose LLM use Lastly, people's first concern shouldn't be watermarking, it should be AI detectors! If you're posting, "its not X, its Y!!", I don't think the watermark is going to make a difference :)
🚨 JUST IN: Claude models will now have invisible watermarks embedded in ALL text, and ALL metadata attached to files…
284
782
252
5,937
1,302,787
Mark Ra retweeted
If you’re not reading the code, whether explicitly or through agentic inquiry, one or more of these is true: ○ You’re a beginner ○ Software is throwaway ○ You’re prototyping ○ You have no users / revenue ○ You’re taking on debt & risk ○ Your problems are basic And btw. All of this is fine. But the reality is that models are still not at the “full autonomy” stage yet. They make rookie mistakes, they go down bad architectural paths. I just had the best model in the world add a nonsensical 700ms delay to “settle” something and it told me “you’re right, I was cargo-culting” 🤨 I am on the camp that this need will diminish more and more. Most code is indeed going to be assembly-like. But we also have the global internet and software infrastructure riding on these models and narrative, and we have to respect that.
570
765
324
8,327
1,329,248
This paper gives a fancy name to a problem you can already feel in your bones "The Tragedy of the Cognitive Commons." Checking AI output requires deep expertise. Deep expertise comes from doing grunt work for years. And grunt work is the first thing AI eats So we're building systems that need expert supervision while dismantling the only known process for making experts The paper calls the shared pool of human expertise the "cognitive commons." Every profession drinks from it. Nobody's refilling it. By eliminating junior roles, each company is acting totally rationally, and the collective result: a profession that can't catch AI's mistakes anymore because it never learned to do the work in the first place Who checks AI's homework in 15 years?
327
3,198
364
12,318
609,332
Mark Ra retweeted
🇭🇷 Croatia got eliminated from the World Cup by a closed-source sensor. FIFA says the Kinexon chip in the Trionda "proved" Matanović touched the ball. Absolute disgrace. What actually happened: they ran signal processing over 500Hz IMU data from a 14-gram sensor and isolated a supposed contact spike from thousands of noisy samples dominated by bladder harmonics, panel flex, spin wobble and stadium micro-tremors, then nullified one of the most dramatic World Cup equalizers in history via spectral analysis of what is essentially glorified noise. Oh, and it gets better. They moved the chip this year. Back in 2022, it used to hang suspended in the center of the ball. Now it's glued into the sidewall of ONE panel, with counterweights stuffed into the other three so the ball doesn't fly like a shopping cart. So sensitivity now depends on which side of the ball the "contact" happens relative to the sensor. Totally fine basis for ending a nation's tournament. An IMU measures acceleration. Somewhere in a proprietary pipeline, a threshold decides which acceleration counts as "touch." That threshold is unpublished. The false positive rate is unpublished. The calibration data is unpublished. The patents are literally still in their secrecy window. And FIFA owns the raw data, so nobody can independently audit the trace that ended Modrić's last World Cup. And we're all supposed to accept a "heartbeat graphic" on the broadcast as if that settles anything. Since we sadly all know you won't ever admit your crimes or remove this technology for good, at least open source the detection pipeline, @FIFAcom! Publish the thresholds, the error rates, the raw IMU trace from last night. If the tech is right, transparency costs nothing, right?
485
2,232
320
14,275
1,259,066
Mark Ra retweeted
Wtf is this commentator on. First she’s acting like Colombia has already won the group in the first round of the group stage. She just called Uzbekistan “Kazakhstan” and multiple times said they were in “Group F”
1
8
370
Mark Ra retweeted
A Brazilian YouTuber killed the Photoshop subscription. It's called PhotoGIMP. It takes GIMP, the free image editor, and makes it look and feel exactly like Photoshop. Same toolbar. Same panel layout. Same keyboard shortcuts. Your hands already know how to use it. Photoshop vs PhotoGIMP: - Price: $275.88 a year → $0 - Account: Adobe login required → No login, ever - Files: Saved to Adobe cloud → Saved on your computer - Updates: Forced when Adobe says → Only when you want - Works on: Windows and Mac → Windows, Mac, and Linux No Adobe account. No cloud upload. No AI trained on your photos. How small is the patch? Tiny. → Nine settings files. That's it. → Copy them into one folder. Done. → Open GIMP. It now looks like Photoshop. → Don't like it? Delete the folder. GIMP goes back to normal. Three steps to install. One command to uninstall. 8,751 stars. 272 forks. 30+ people from around the world helping translate it. One honest note: the license is GPL-3.0. Free for everything. Personal work, paid client work, your own edits. No "Pro" tier hiding behind it. Dionatan Simioni runs the biggest Linux YouTube channel in Brazil. He built this from Marau, a small town in Rio Grande do Sul. No VC. No team. No fundraise. This is what Photoshop should have been from the start. (Link in the comments)
196
2,140
112
13,501
685,535
Replying to @ChampRDS
I dont understand how you give a decision in favor of the challenger in that fight. At the very least, it was a draw and the champ should retain. All Strickland did is defend the takedown very well for one round and throw jabs. The rest was neutral or he was controlled.
1
5
983
A mathematician who shared an office with Claude Shannon at Bell Labs gave one lecture in 1986 that explains why some people win Nobel Prizes and other equally smart people spend their whole lives doing forgettable work. His name was Richard Hamming. He won the Turing Award. He invented error-correcting codes that made modern computing possible. And he spent 30 years at Bell Labs sitting in a cafeteria at lunch watching which scientists became legendary and which ones faded into nothing. In March 1986, he walked into a Bellcore auditorium in front of 200 researchers and told them exactly what he had seen. Here's the framework that has been quoted by every serious scientist for the last 40 years. His opening line landed like a punch. He said most scientists he worked with at Bell Labs were just as smart as the Nobel Prize winners. Just as hardworking. Just as credentialed. And yet at the end of a 40-year career, one group had changed entire fields and the other group was forgotten by the time they retired. He wanted to know what the difference actually was. And he said it wasn't luck. It wasn't IQ. It was a specific set of habits that almost nobody is willing to follow. The first habit was the one that hurts the most to hear. He said most scientists deliberately avoid the most important problem in their field because the odds of failure are too high. They pick a safe adjacent problem, solve it cleanly, publish it, and move on. And because they never swing at the hard problem, they never hit it. He said if you do not work on an important problem, it is unlikely you will do important work. That is not a motivational line. That is a logical one. The second habit was about doors. Literal doors. He noticed that the scientists at Bell Labs who kept their office doors closed got more done in the short term because they had no interruptions. But the scientists who kept their doors open got more done over a career. The open-door scientists were interrupted constantly. They also absorbed every new idea passing through the hallway. Ten years in, they were working on problems the closed-door scientists did not even know existed. The third habit was inversion. When Bell Labs refused to give him the team of programmers he wanted, Hamming sat with the rejection for weeks. Then he flipped the question. Instead of asking for programmers to write the programs, he asked why machines could not write the programs themselves. That single inversion pushed him into the frontier of computer science. He said the pattern repeats everywhere. What looks like a defect, if you flip it correctly, becomes the exact thing that pushes you ahead of everyone else. The fourth habit was the one that hit me the hardest. He said knowledge and productivity compound like interest. Someone who works 10 percent harder than you does not produce 10 percent more over a career. They produce twice as much. The gap doesn't add. It multiplies. And it compounds silently for years before anyone notices. He finished the lecture with a line I have never been able to shake. He said Pasteur's famous quote is right. Luck favors the prepared mind. But he meant it literally. You don't hope for luck. You engineer the conditions where luck can land on you. Open doors. Important problems. Inverted questions. Compounded hours. Those are not traits. Those are choices you make every single day. The transcript has been sitting on the University of Virginia's computer science website for almost 30 years. The video is free on YouTube. Stripe Press reprinted the full lectures as a book in 2020 and Bret Victor wrote the foreword. Hamming died in 1998. He gave his final lecture a few weeks before. He was 82. The lecture that explains why some careers become legendary and others disappear is still free. Most people who could benefit from it will never open it.
143
1,835
183
8,083
1,183,369
Mark Ra retweeted
🚨Architects are going to hate this. Someone just open sourced a full 3D building editor that runs entirely in your browser. No AutoCAD. No Revit. No $5,000/year licenses. It's called Pascal Editor. Built with React Three Fiber and WebGPU -- meaning it renders directly on your GPU at near-native speed. Here's what's inside this thing: → A full building/level/wall/zone hierarchy you can edit in real time → An ECS-style architecture where every object updates through GPU-powered systems → Zustand state management with full undo/redo built in → Next.js frontend so it deploys as a web app, not a desktop install → Dirty node tracking -- only re-renders what changed, not the whole scene Here's the wildest part: You can stack, explode, or solo individual building levels. Select a zone, drag a wall, reshape a slab -- all in 3D, all in the browser. Architecture firms pay $50K+ per seat for BIM software that does this workflow. This is free. 100% Open Source.
174
957
48
5,871
593,363
Mark Ra retweeted
Why is no one talking about this? @nvidia is offering around 80 AI models via hosted APIs absolutely for free. You get access to MiniMax M2.7, GLM 5.1, Kimi 2.5, DeepSeek 3.2, GPT-OSS-120B, Sarvam-M etc. This plugs straight into OpenClaude, OpenCode, Zed IDE, Hermes agent and even with Cursor IDE. Setup: – Grab API key: build.nvidia.com/models – base_url = "integrate.api.nvidia.com/v1" – api_key = "$NVIDIA_API_KEY" – select model (e.g. minimaxai/minimax-m2.7) If you’re building or experimenting, this is basically free inference. Lock in and start building today anon. Thank me later.
531
1,826
134
18,114
1,698,211