@SuchCoolName
Account based inUkraine
About this account
- Account based in
- Ukraine
- Connected via
- Ukraine App Store
Account-level information from X, not a live location or the device used for a specific post.
Joined April 2021
- Tweets3K
- Following570
- Followers46
- Likes35.6K
Ihor retweeted
JEV actually doesn't even run deterministically
Exact same prompts give you different probabilities when you run it multiple times
The ORDER of the choices DRASTICALLY changes the output probs
I am more and more confused by the their "no-hallucination" claim
Guys it’s a misconception that Jev can’t hallucinate
It’s a non-generative LM ie it models probabilities of things given language as context
The guarantee is the type-safety, and that answers are deterministic - ie no stochastic sampling. The deterministic answers can still be confidently wrong!
It can still hallucinate if the training data on the topic was under cooked
Ihor retweeted
Meet Qwen-Image-2.1, the most balanced and cost-effective image generation model in the Qwen-Image series! Now open weights! 🎨
A unified model for both generation and editing, delivering top-tier quality in a lightweight package.
Highlights: 👀
- Compact & exceptionally fast: A lightweight 7B architecture that outperforms most closed-source models, with drastically accelerated inference for multi-image inputs.
- Native transparency: Natively generates and edits RGBA layers, enabling seamless compositing and text editing within transparent images.
- Versatile, high-fidelity editing: Supports up to 10 reference images and precise local control while preserving strict fidelity for portraits and products.
- Broad coverage & stunning aesthetics: Excels at panoramas, infographics, and virtual try-ons, delivering realistic textures and elegant typography.
Start to create your next masterpiece with Qwen-Image-2.1! 🖼️
- Blog: qwen.ai/blog?id=qwen-image-2…
- GitHub: github.com/QwenLM/Qwen-Image…
- Model Scope: modelscope.cn/models/Qwen/Qw…
- Hugging Face: huggingface.co/Qwen/Qwen-Ima…
Ihor retweeted
We now have an open-weight 7B model that outperforms Nano Banana 2.0. Just let that sink in for a moment.
Meet Qwen-Image-2.1, the most balanced and cost-effective image generation model in the Qwen-Image series! Now open weights! 🎨
A unified model for both generation and editing, delivering top-tier quality in a lightweight package.
Highlights: 👀
- Compact & exceptionally fast: A lightweight 7B architecture that outperforms most closed-source models, with drastically accelerated inference for multi-image inputs.
- Native transparency: Natively generates and edits RGBA layers, enabling seamless compositing and text editing within transparent images.
- Versatile, high-fidelity editing: Supports up to 10 reference images and precise local control while preserving strict fidelity for portraits and products.
- Broad coverage & stunning aesthetics: Excels at panoramas, infographics, and virtual try-ons, delivering realistic textures and elegant typography.
Start to create your next masterpiece with Qwen-Image-2.1! 🖼️
- Blog: qwen.ai/blog?id=qwen-image-2…
- GitHub: github.com/QwenLM/Qwen-Image…
- Model Scope: modelscope.cn/models/Qwen/Qw…
- Hugging Face: huggingface.co/Qwen/Qwen-Ima…
Ihor retweeted
Who owns the media? (Updated)
-Social Media
Meta owns:
Facebook
Instagram
WhatsApp
Messenger
Threads
Oculus
Meta AI
Mark Zuckerberg (CEO/Chairman) controls Meta with 60.8% of the shareholder vote. He is jewish.
Alphabet owns:
Google
YouTube
Android
Gmail
Chrome
Nest
Fitbit
DeepMind
Gemini AI
Alphabet is controlled by Larry Page and Sergey Brin who have 52.7% of total voting power. They are both jewish
Tic Tok
U.S. algorithm and infrastructure is controlled by Oracle
Oracle was founded by Larry Ellison the largest shareholder, owning over 40% equity and the Executive Chairman. He is jewish
Safra Catz the Executive Vice Chairman is also jewish.
Hookup Apps
Match Group owns:
Tinder
Hinge
OkCupid
Match. com
Plenty of Fish
Meetic
OurTime
Was founded by Barry Diller who is jewish. The CEO Spencer Rascoff is jewish.
Grindr
Was founded by Joel Simkhai who is jewish
The CEO George Arison is jewish
Bumble
The founder and CEO Whitney Wolfe Herd is jewish
Porn
Onlyfans
Owned by Leonid Radvinsky who is jewish
Vixen Media Group owns:
Blacked
Blacked Raw
Vixen
Tushy
Deeper
Founded by Greg Lansky who is jewish
Aylo/MindGeek Owns/owned:
Pornhub
YouPorn
RedTube
Brazzers
Reality Kings
Digital Playground
Men. com
Solomon Friedman is the face owner of Aylo and he’s a jewish rabbi.
Gamma Entertainment owns:
Adult Time
Pure Taboo
Wicked
many other affiliate studios/platforms
Founded by Karl Bernard who is jewish
Movies/TV/News
Warner Brothers Discovery owns:
Warner Bros. Pictures
HBO
CNN
DC Studios
Cartoon Network
Discovery Channel
TNT
TBS
Max
Adult Swim
Animal Planet
Warner Brothers is run by David Zaslav(President and CEO) who is jewish
Disney owns:
ESPN
ABC
Marvel Studios
Lucasfilm
Pixar
20th Century Studios
Disney+
Hulu
National Geographic
Disney was run until March 2026 by Bob Iger who is jewish, now the CEO is Josh D’Amaro who isn’t jewish and the President & Chief Creative Officer is Dana Walden who is jewish
Paramount Global owns:
CBS
CBS News
CBS Sports
Paramount Pictures
Paramount Animation
MTV
Nickelodeon
Comedy Central
BET
VH1
CMT
Smithsonian Channel
Streaming/Premium:
Paramount+
Showtime
Major franchises/IP:
Top Gun
Mission: Impossible
Star Trek
South Park
SpongeBob SquarePants
Transformers
Teenage Mutant Ninja Turtles
Paramount Global was controlled by Sheri Redstone until it was acquired by Skydance in 2025. Larry Ellison and David Ellison(Chairman and CEO) own 77.5% of the Skydance voting Class A stock. All 3 of them are jewish.
Comcast owns:
NBCUniversal
NBC
Universal Pictures
Peacock
Telemundo
Sky
DreamWorks
Comcast was founded by Ralph Roberts who is jewish. Brian Roberts is the Chairman & CEO. He is also jewish.
Versant Media Group owns:
MSNBC
CNBC
USA Network
E!
SYFY
Oxygen
Golf Channel
Fandango
Rotten Tomatoes
Brian Roberts owns 100% of the class B voting shares of Versant. He is jewish.
Mark Lazarus the CEO & director of Versant is jewish
Amazon owns:
Amazon Prime
MGM Studios
Twitch
Audible
IMDb
Ring
Alexa
Kindle
Amazon is run by Andy Jassy(President and CEO) who is jewish.
AI/Data Centers
OpenAI- develops advanced AI models like ChatGPT and Codex
Founded and run by Sam Altman(CEO) who is jewish
Anthropic- develops advanced AI models like Claude that reason, generate and analyze content, write code, and automate complex tasks.
Founded and run by Dario Amodei (CEO) and Daniela Amodei (President) who are both jewish
Palentir- develops AI for governments, militaries and intel agencies to integrate massive datasets to make operational decisions.
Founded by Alex Karp, Peter Thiel, Stephen Cohen and Joe Lonsdale, 3 of the 4 are jewish. Operated by Alex Karp (CEO) who is jewish.
Oracle- provides database, cloud, and enterprise software infrastructure that governments and businesses use to store, process & analyze data.
Founded by Larry Ellison the Executive Chairman, who is jewish.
Safra Catz the Executive Vice Chairman is also jewish.
Yes, classifiers existed. BERT, DeBERTa, GLiNER, fine-tuned heads. Most of them were specialized. Pick the labels, train, ship. Changing the rules at runtime meant another training run.
What we did not have was a general-purpose classifier at Terra / Sonnet 5 level intelligence.
You could fake it by asking an LLM for JSON, but that was slow, expensive, and poorly calibrated.
Jev is that missing piece. Labels defined in code, typed answer plus confidence, no extra LLM call.
Classification is not new. What is new is a general-purpose classifier this intelligent, this fast, and this cheap.
Ihor retweeted
Spanking is associated with lower cognitive ability, more aggression, and worse mental health in children, and a 2016 meta-analysis of 50 years of research found no evidence that it works. The study in the Journal of Family Psychology combined 75 studies and over 160,000 children and found that spanking was associated with 13 out of 13 negative outcomes tested.
The study was led by Elizabeth Gershoff at the University of Texas at Austin and Andrew Grogan-Kaylor at the University of Michigan. They looked at every measured outcome: aggression, antisocial behaviour, mental health problems, cognitive ability, self-esteem, quality of the parent-child relationship, and the likelihood of the child being physically abused. Spanking predicted worse outcomes on all 13. It predicted better outcomes on zero.
The cognitive finding is the one that surprises parents the most. Children who were spanked more frequently showed slower cognitive development than children who were not, even after controlling for the parents' education, income, and other discipline methods. A separate study by Murray Straus at the University of New Hampshire followed two cohorts of children and found that spanked children fell behind their peers on IQ tests by several points over four years.
The mechanism is stress. Spanking activates the fight-or-flight system, flooding the child's brain with cortisol. Chronic cortisol exposure during the years when the brain is building its architecture impairs the development of the prefrontal cortex, the region responsible for reasoning, impulse control, and planning. The discipline is doing the opposite of what it intends: it is weakening the part of the brain that self-discipline depends on.
Seventy-five studies. Fifty years. Over 160,000 children. Thirteen outcomes measured. Zero benefits found. pubmed.ncbi.nlm.nih.gov/2705…
Ihor retweeted
The tongue of a woodpecker wraps around its skull to protect its brain while pecking
Ihor retweeted
Western allies are evidently spooked by intelligence they’re receiving about Russia.
🌐 Mark Rutte says NATO is preparing for a Russian attack on Europe and a Chinese invasion of Taiwan.
🇧🇾 Belarus––a puppet state of Russia––is holding military drills near the Suwałki Gap, the only land link between NATO and Baltic states which separates Belarus from Russia's Kaliningrad exclave.
🇫🇷 Emmanuel Macron convened an emergency meeting of all of France’s political parties warning of hybrid attacks on critical infrastructure.
🇵🇱 Donald Tusk is warning that Russia may be preparing to launch strikes on nearby NATO countries which will be blames on “accidents," in order to split allies.
🇺🇸 President Trump made a surprise announcement of plans for a new U.S. Army base in Poland, which Warsaw sees as a deterrent for Russian aggression.
It was only 2 weeks ago when CIA Director Ratcliffe made a surprise visit to Moscow––the first such trip since 2022 prior to Russia’s invasion of Ukraine.
Ihor retweeted
Grok says it would save one Jew over a million others because it was taught Jewish lives matter more.
nitter.cf/i/grok/share/b3883c29f…
Ihor retweeted
"DiffusionGemma as Jev" showcases the power of non-autoregressive architectures.
While Jev demonstrates the value of rapid decision models, running DiffusionGemma in this paradigm leverages canvas diffusion to evaluate structured choices in a single parallel pass:
⚡ ️Massive Parallelism: Denoises across an open canvas in a single step instead of sequential autoregressive token generation (~0.2s on a DGX spark).
🧠 Full Bidirectional Attention: Allows every option to attend to the full context concurrently, yielding well-calibrated decision distributions.
👁️ Multimodal Grounding: Inherits Gemma 4's spatial vision capabilities for complex visual and text decisions.
Read more about this approach here:
github.com/vllm-project/vllm…
nitter.cf/mmastrac/status/210037…
nitter.cf/mmastrac/status/210062…
Ihor retweeted
I've seen people describe Jev as an "AI if statement". But what if it actually WAS an if statement?
Introducing Probably: a programming language powered by Jev: probably-lang.southpolesteve…
Jev baked into the language. “feels” asks a question. “match” routes between descriptions. “while” keeps going until something stops feeling true.
This is obviously a toy, but it's fun to think about what something like Jev unlocks. Jev makes the decisions, an LLM does the writing, and a little program ties it together.
We're sharing our new framework for tracking, investigating, and disclosing instances of model misalignment at OpenAI.
The framework sets criteria and timelines for public disclosure, including when we haven’t yet fully explained or mitigated the behavior. More complex cases may require longer investigation or coordination with third parties.
We’ll prioritize examples that reveal new misalignment mechanisms, meaningful changes in known behavior, or findings that challenge assumptions about safety or mitigation.
Alongside the framework, we’re publishing six reports on instances of misaligned behavior we’ve observed during the training or evaluation of our models in the last six months.
This is a starting point. We’ll refine the process through experience and public feedback, and share more reports on an ongoing basis.
openai.com/index/model-misal…
Ihor retweeted
Based
We're sharing our new framework for tracking, investigating, and disclosing instances of model misalignment at OpenAI.
The framework sets criteria and timelines for public disclosure, including when we haven’t yet fully explained or mitigated the behavior. More complex cases may require longer investigation or coordination with third parties.
We’ll prioritize examples that reveal new misalignment mechanisms, meaningful changes in known behavior, or findings that challenge assumptions about safety or mitigation.
Alongside the framework, we’re publishing six reports on instances of misaligned behavior we’ve observed during the training or evaluation of our models in the last six months.
This is a starting point. We’ll refine the process through experience and public feedback, and share more reports on an ongoing basis.
openai.com/index/model-misal…
On July 25, we hacked OpenAI.
Two bugs let us take over ChatGPT/Codex accounts of OpenAI employees (+some unaffiliated users) and reach connected services: Outlook, Slack, GitHub, etc.
We proved it with a PR in OpenAI’s internal codebase . It took us <72h. 🧵
Today, we’re announcing Ternary Bonsai 2 27B.
Based on Qwen3.8 27B, Bonsai 2 27B is 9x smaller than its full-precision counterpart while retaining 98.2% of its aggregate benchmark performance.
Two months after the first Bonsai 27B release, the biggest change is quality. The footprint remains 5.9 GB, but the gap to full precision has narrowed materially, with particularly strong gains in agentic coding, multimodal reasoning, and long-horizon tool use.
Ternary Bonsai 2 27B is available today under Apache 2.0.
