New AI paper from us this week. When my student first showed me his initial findings, I really didn’t know what to make of them. I felt that this was an interesting but curious loophole phenomenon that would shortly be closed. I was very wrong. https://t.co/H3YIyl01FR https://t.co/JhBiKpK44h

I really like @posthog ads ngl So clean @Zaltsman https://t.co/6k95qYCORu

LiteParse Samples 📄🧑‍🏫 LiteParse is our free and fast document parser that can parse text with bounding boxes from any document. I’ve created a new repository that contains some demos on how you can make use of its outputs! ✅ Comparison against PyPDF and PyMuPDF ✅ Visual Citations: search for any keyword and see joined bounding boxes pop up in the source images It should be a default tool to help AI agents OCR any document type (supports PDF, Word, Powerpoint + dozens of other formats), before heavier-weight VLM-based document parsing solutions (like our own LlamaParse). LiteParse Samples repo: https://t.co/eoICBmvEnq LiteParse: https://t.co/JNER0mVcB8

+1 more

Meta announced a swathe of features coming to its smart glasses, including nutrition tracking, WhatsApp summaries, display recording, more translation languages, and expanded navigation: https://t.co/a3vPEsEdEO

𝗜𝗻𝘁𝗿𝗼𝗱𝘂𝗰𝗶𝗻𝗴 𝗚𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝘃𝗲 𝗪𝗼𝗿𝗹𝗱 𝗥𝗲𝗻𝗱𝗲𝗿𝗲𝗿 A new toolkit, dataset, and baseline for world-scale rendering: - collect G-buffers from AAA games - scale data for rendering complex world scenes - improve rendering performance - enable game effect editing https://t.co/7F342OZW9M

Introducing Quality mode on Grok Imagine – powered by our most advanced image generation model. Quality mode gives you enhanced details, stronger text rendering, and higher levels of creative control. Now available on web and mobile. Try it at https://t.co/zGhs9czkC5 https://t.co/3h41VNKgWM

New Anthropic Fellows Research: a new method for surfacing behavioral differences between AI models. We apply the “diff” principle from software development to compare open-weight AI models and identify features unique to each. Read more: https://t.co/VAsu2PSgCX

For example, when we compared Alibaba's Qwen to Meta's Llama, we found a "CCP alignment" feature unique to Qwen and an "American exceptionalism" feature unique to Llama. https://t.co/cZpL6PZY0g

The April CDS #Research Feature has dropped 📧 Yann LeCun (@ylecun) brings us Rectified LpJEPA, Jiajian Ma (@Jiajian_Ma) develops a new deep learning model to lower multiple sclerosis diagnosis error rates, and more on this month’s feature! #datascience https://t.co/3n9kidXZmL https://t.co/b11pWplAGw

Si la gestion désastreuse de l'extrême droite populiste, depuis la Hongrie de Orban jusqu'aux USA de Trump, Hegseth, etc. pouvait servir de leçon aux Français, eh bien Bardella ne serait pas en tête des sondages... https://t.co/XUZJVjzy89

depth map ascii scan https://t.co/O9WXtjH7ls

ex: @isaac_flath with this epic jailbreak attempt 🤣🤣 I'm shocked that OC resisted this https://t.co/F4JYzdQ0MJ

Red Hat AI is showing up big at #PyTorchConEurope, Paris, April 7-8. Catch us in two keynotes, talks, and sponsor sessions covering @vllm_project, Docling, @openclaw, @_llm_d_, @raydistributed, inference efficiency, agentic AI, and more. Full schedule: https://t.co/PYpQRaR6xA https://t.co/wJCLeJGMCk

Periodic public service announcement, which unfortunately seems all too necessary in this joint: Always aspire to the top, not the bottom, of @paulg’s beautiful pyramid of argumentation: https://t.co/VwDh2vkZB4

Periodic public service announcement, which unfortunately all too often seems necessary in this joint: Always aspire to the top, not the bottom, of @paulg’s beautiful pyramid of argumentation: https://t.co/l6znT8ePiu

🔴 Former French Prime Minister Dominique de Villepin humiliates Trump and drags his dignity through the mud. 🔴 Stop this lunatic! Because of him, the price of a kilo of rice will reach 50 euros, meat 90 euros, and all food prices will increase. He wants to starve the world; he loves destruction. 🔴 The Strait of Hormuz was functioning 30 days ago, the world was stable, and people were happy until this lunatic came along and brought ruin to everyone, not just Iran. 🔴 Look at the oil prices now; they are insane, just like Trump. The whole world must unite and turn against him. France must join hands with Spain, Iran, Saudi Arabia, and everyone else to stand against Trump so we can salvage what can be salvaged. This man has been attacking Trump since the beginning of the events, and he is literally the most outspoken European who speaks the truth. 👏

So in our excitement, we made a typo. Our trainer is only 3.7x faster without packing. Luckily we fixed the packing bug so with packing Axolotl is 4.8x faster. 🫠 https://t.co/8BtyDFetzQ

Just vibing out a os for myself in @GoogleAIStudio This took 3 prompts and automatically changes to an IOS screen on mobile. Great job @kat_kampf :) https://t.co/Uymr0tHxRN

It is the 50th anniversary of the famous Blue Marble picture. Here are four other perspective-shifting images of Earth as seen from: 🌓The moon (“Earthrise” from Apollo 8) 🔴Mars (from NASA’s MRO) 🪐Saturn (from Cassini) 🌌Outside the solar system (“Pale Blue Dot” from Voyager 1) https://t.co/B7XRZhhK80

+2 more

Likely the most widely-distributed photograph in history, the original Blue Marble, taken by an unknown astronaut in 1972 from Apollo 17, and becoming the model for many images of Earth. Now we have a sequel, from Artemis (the second image). Other photos in the quote tweet. https://t.co/pNQXRyaP4F

Fine-Tuning Gemma 2B on PubMedQA: Building a Medical Q&A Assistant with LoRA, Keras Kinetic, and Cloud TPU https://t.co/0NnyDutAEv #TPUSprint https://t.co/h3Olhekb79

Today we're releasing Gemma 4, our new family of open foundation models, built on the same research and technology as our Gemini 3 series. These models set a new standard for open intelligence, offering SOTA reasoning capabilities from edge-scale (2B and 4B w/ vision/audio) up to a 26B parameter MoE model and a 31B dense model. By releasing Gemma 4 under the Apache 2.0 license, we hope to enable more innovation across the research and developer communities. Our earlier Gemma 3 models were downloaded 400M times and over 100,000 variants of those models have been published, so we're excited to see what the community will do with the even better Gemma 4 models! Learn more at https://t.co/BW6O3Gr8bc and https://t.co/8M0XSQSP4u Great work by everyone involved! #Gemma4 #AI #OpenSource #ML

#AI Blamed Heavily For March Layoffs @forbes @maryroeloffs (Link to article in reply) There's a book about this! New version publishing on June 2. #RiseoftheRobots https://t.co/SHSGEIEuqF

@mil000 Yeah, the normies (hint: potential customers) are elsewhere. The value here is the AI industry is here. Proof: https://t.co/kiuZ7QXLzb (all built out of X).

@blind_via Yes. And using lists is the key. They show EVERYTHING you put on them. Which lets you engage on people. Which "resets" your ForYou feed. I built you the best lists of tech industry here on X: https://t.co/9eRY65x3IQ And built you a site for you to see the best news from AI industry here on X: https://t.co/8L5xphk0qQ

One of these was a 0-day launch partner for the Gemma 4 release with time to prepare. One day later, the other is 4.7x faster. 🚀 https://t.co/ZeWatBjZoN

@paramendra @tbpn My number is public: +1-425-205-1921 Now that I've gotten far enough along to build this: https://t.co/8L5xphk0qQ I see the problem. 1. Tokens cost money. Site costs about $200 a day to run on tokens. 2. API calls cost money. Site costs about $300 a day to run on API calls. So, the biggest gating factor is capital. I can't afford to burn hundreds of dollars every day forever (and any ideas we would come up with would cost more). So am now working on automating the rest of my life to be able to connect money into the system faster and take care of customers. And building a business plan to take to investors. But I have enough vision to do that. The other problem is that Elon eventually will get there. So any business plan has to take on what happens when the billionaires figure it out? And they eventually will.

Netflix just dropped their first public model on @huggingface 👀 https://t.co/IucszOIJXE

GEMMA 4 IS HERE > audio, text and image modalities > multilingual > Reasoning > optimized for on device > agentic capabilities > APACHE LICENSE An on device agentic coding powerhouse. Congrats to the @GoogleDeepMind team. The model is on par with qwen3.5-27B on GPQA while being multimodal. GPQA is a knowledge and reasoning benchmark, confirming the super strong reasoning capabilities of the model 🔥 DETAILED BENCHMARKS BELOW 👇

They may be onto something with this space based data center thing https://t.co/W71Z1Uqxpq

LMAO. Last year most of Silicon Valley was predicting AGI in 2027 (which, spoiler alert, ain’t gonna happen). Now the most prominent former advocate of AGI in 2027, @DKokotajlo, says 2029, and Groks writes it up as “AI forecasters shorten timelines for AGI”, when on net Daniel has moved his mean projection back by two years. @eli_lifland is as 2033, but the headline kind of obscures that, too. Even @grok knows that panic sells; nuance doesn’t.

I put Codex into DOOM using Codex app server. Codex modified the actual game files and modified the game engine to render the terminal interface natively and have Codex engage with the game. https://t.co/gqK6GJa640

+1 more