This is a terrible user experience. Surely there must be a better way? Perhaps a reminder of just how early we are in the AI game. https://x.com/bilawalsidhu/status/1887363382883356982
Let’s goo! Google presents Scaling Pre-training to One Hundred Billion Data for Vision Language Models results highlight that while traditional benchmarks may not benefit significantly from scaling noisy, raw web data to 100 billion examples, this data scale is vital for https://x.com/_akhaliq/status/1889526316673732753
Google presents: Scaling Pre-training to One Hundred Billion Data for Vision Language Models – Studies pre-training VLMs on an unprecedented scale: 100B examples! – Perf saturates at this scale on many common Western-centric benchmarks. – Tasks of cultural diversity and https://x.com/arankomatsuzaki/status/1889520451501900000
Google released Gemini 2.0 Flash Thinking Experimental 1-21, the latest version of its vision-language reasoning model, with an expanded 1 million-token context window and a user-readable chain of thought. The update improves accuracy across science, math, and multimedia https://x.com/DeepLearningAI/status/1889026549275344986
Scaling Pre-training to One Hundred Billion Data for Vision Language Models Google DeepMind develops WebLI-100B, a novel dataset containing 100 billion image-text pairs, representing a tenfold increase over the largest reported vision-langauge datasets. They demonstrate that a https://x.com/iScienceLuvr/status/1889521157482959067
New ways to help you drive performance with Demand Gen https://blog.google/products/ads-commerce/new-demand-gen-features-2025/
Hot take: imagen 3 is now better than midjourney and flux Give it a try in gemini app or FAL — it’s a ton of fun https://x.com/bilawalsidhu/status/1889147763297812672
It takes less than 30 seconds to create @GoogleDeepMind Gemini API key and start using Gemini 2.0 Flash! 👀 1️⃣ Have a Google account (no Google Cloud required) 2️⃣ Go to AI studio “aistudio(.)google(.)com” 3️⃣ Click on “Get API Key” top-left corner and “create API Key” 4️⃣ Send https://x.com/_philschmid/status/1889689838464516228
Deep Research comes to Gemini Android app https://9to5google.com/2025/02/04/gemini-deep-research-android/
Google removes pledge to not use AI for weapons from website | TechCrunch https://techcrunch.com/2025/02/04/google-removes-pledge-to-not-use-ai-for-weapons-from-website/
🤔 An Energy Star for AI? Introducing AI Energy Score: First-ever rating system comparing 166 AI models’ energy consumption! From LLaMa to Gemma, get transparent ⭐️1-5 efficiency ratings. Incredible work led by @SashaMTL https://x.com/fdaudens/status/1889248951339024552
I think about all the companies building around small, less capable models because they were cheap months ago & then Gemini 2.0 Flash comes along and you get very smart and fast and cheap. Really illustrates why you shouldn’t bet against the improvement curve in AI anytime soo. https://x.com/emollick/status/1887558394522824940
Responsible AI: Our 2024 report and ongoing work https://blog.google/technology/ai/responsible-ai-2024-report-ongoing-work/
Google I/O 2025: Save the date https://blog.google/feed/google-io-2025-save-the-date/
AI and the future of scientific leadership https://blog.google/technology/ai/ai-future-of-scientific-leadership/
man, still a long way to go to run down google 🥺 / X https://x.com/sama/status/1888703820596977684
Google presents Gold-medalist Performance in Solving Olympiad Geometry with AlphaGeometry2 https://x.com/_akhaliq/status/1887718062863855625
Today we’re expanding the Gemini 2.0 family with new options and broader availability. This builds on the first model we launched in December: 2.0 Flash, our model with low latency and better performance ⚡ Read more on today’s launches ⬇️ https://x.com/Google/status/1887170927751729385
Gemini 2.0 model updates: 2.0 Flash, Flash-Lite, Pro Experimental https://blog.google/technology/google-deepmind/gemini-model-updates-february-2025/
Of the three models showing their “thinking,” Gemini may actually be the most interesting. Less personal and cute than DeepSeek r1, not obscured like the OpenAI models, and actually shows the pathways it took to solve the problem more clearly, making it easier to diagnose issues. https://x.com/emollick/status/1887979499100512535
Not an easy chart to read, but worth spending a minute on. It tells you a few key things: 1) Gemini 2.0 Flash Thinking sets a new high in price-performance, better than DeepSeek r1 (on ELO) and cheaper 2) The cost of GPT-4 capability dropped 1000 fold in 18 months 3) Pace is fast” / X https://x.com/emollick/status/1887401985688662509
@ZyphraAI have launched their first Text to Speech model, Zonos-v0.1 – now the leading open weights Text to Speech model in Artificial Analysis Speech Arena. Zonos-v0.1 is currently scoring a Speech Arena ELO of 1020, just behind proprietary speech models available on Google https://x.com/ArtificialAnlys/status/1889150365913972930
Google announces new partnership to prevent and treat women’s cancers https://blog.google/technology/health/google-ai-institute-womens-cancers/
Gemini 2.0 Pro with: “create something I can paste into p5js that will startle me with its cleverness in creating something that invokes the control panel of a starship in the distant future” the first time (Gemini 1.5 Pro couldn’t do it, other examples from other models below) https://x.com/emollick/status/1887342222560870629
The discussion about Deep Research shows the gaps between fields for what they think “research” is. OpenAI Deep Research is not great at digging up tons of citations or facts (Google’s is better at that) Deep Research is great at digging up arguments, holes & providing analysis https://x.com/emollick/status/1887295987284074580
Seems like increasing evidence that LLMs that are not Reasoners are going to fade away. There’s the explicit announcement from OpenAI, the academic studies on the topic, the fact that Gemini 2.0 Pro (and apparently Grok 3) seem to be outclassed by their reasoning AI cousins, etc.” / X https://x.com/emollick/status/1889802272454090921
Roblox Joins Alphabet, OpenAI in ROOST Fund to Boost Child Safety Online – Bloomberg https://www.bloomberg.com/news/articles/2025-02-10/roblox-joins-tech-companies-in-new-fund-to-protect-kids-online
Three models now pass my Lem Test on the first try: Claude 3.5 Sonnet, o1-pro and now Gemini 2.0 Flash Thinking (though o1-pro has the strongest verse and Flash Thinking the weakest) Everything else fails, including DeepSeek r1, o3-mini-high, and Gemini 2.0 Pro https://x.com/emollick/status/1887348169622565018
“More people watch YouTube on TV than they do on mobile and computers. YouTube is quite literally the new TV. https://x.com/bilawalsidhu/status/1889709040634106059
