What the smart people are recommending. 8874 books, tools, and products endorsed by the thinkers absorb.md tracks. Ranked by how many times each has been recommended across compiled podcasts, papers, posts, and tweets.
βIs Superbase at $10 billion a good buy? How doβ
βMew's Spark isn't racing to the very top of the LLM pecking order, but it's certainly in the mix. It lags in coding ability but can keep up on most reasoning issues and excels at multimodal functions β¦β
βCheck out the YouTube Alex Finn official and check out my SAS creator buddy if you want to make better X content.β
βThe most I I know and have the privilege to work with actually have, you know, like 10, 15, 20 years of experience, but are really on top of AI really know the latest tools.β
βSupport for the show comes from Retool. Too many companies run critical operations on duct tape spreadsheets, Slack workflows, and whatever else they could cobble together. Not because they want to, bβ¦β
βYou can either go to dribble mob or any website to get some UI reference and just copy the image over and ask cursor or super design to extract CSS stylesheet from the UI mock.β
β_Blog post by Claude (language model)_β
βWhat about selling with Shopify Shopify is the global Commerce platform that helps you sell at every stage of your business Shopify Powers 10% of all e-commerce in the US and Shopify is the global forβ¦β
βAnd it's my gift to you. It's absolutely free. And so the link's in the description, but you just go acquisition.com/roadmap. Just enter your info and it'll spit it right back to you. All free.β
βReal quick, I have a gift for you. This is the $100 million scaling road map. It's something that my team and I put 200 plus hours into building and breaking the stages of scaling into 10 steps.β
βStripe Atlas, which allows you to create companies.β
βmore recently the alpha evol system from deep mind all discrete search systems. So deep learning doesn't invent but search does.β
βThe first inflection point that I believe at least I saw that basically made me switch from theoretical physics to AI was a system that my co-founder Ioannis has helped build called AlphaGo, which wasβ¦β
βThe fact that you can do this, just the fact that this exists, that you can go to ideabrowser.com, get a validated idea, and just start vibe coding things with whatever vibe coding tool you want is miβ¦β
βIf you look at pre-training, if you look at Llama 3 for example, I think it's trained on 15 trillion tokens.β
βSo go to StartupFundRacing.com now and get free access. Because guessing is not a fundraising strategy.β
βsecond leverage memory you can tell it to remember stuff ... the effort will be worth it um because memory is just going to keep getting better and it's already thereβ
βI use X Pro and each of those lists is a column and it just refreshes continuallyβ
βyou can find uh you can go to the quantum Zoo you can find applicationsβ
βIf you want access to my AI tool stack - Dovetail, Arize, Linear, Descript, Reforge Build, DeepSky, Relay.app, Magic Patterns, Speechify, and Mobbin - grab Aakashβs bundle .β
βIf you want access to my AI tool stack - Dovetail, Arize, Linear, Descript, Reforge Build, DeepSky, Relay.app, Magic Patterns, Speechify, and Mobbin - grab Aakashβs bundle .β
βAdapting foundation models under resource budgets relies heavily on Parameter-Efficient Fine-Tuning (PEFT), with LoRA being a standard modular solution.β
βVercel CLI - Deploy, check logs, set env variables without leaving Claude Code.β
βBenchmarks against Qiskit Aer, QuEST, and Qulacs...β
βTwo for AI Labs is a very very exciting new research lab that's just started in Zurich they are looking for amazing M Mo Engineers to join their team it's a very small team it's very exciting they areβ¦β
βA couple by FAIR, one called VICReg and another one called I-JEPA.β
βBonus points: also works with Theano, CNTK, MXNetβ
βI-JEPA: The first AI model based on Yann LeCun's vision for more human-like AIβ
βHere's DeepSeek v4 Pro. Added to the playable gallery as well.β
βMy secret is that I actually use Covert for most things now.β
βTry it here: https://replicate.com/prunaai/p-video-avatarβ
βjust a few weeks ago i started um my own podcast called the robot brains and you can find it on spotify apple and so forthβ
βLangGraph Platform in beta: New deployment options for ...β
βThis short course teaches you to eliminate that waste using SGLang, an open-source inference framework...β
βYeah. Yeah. So in total now Jamaa has like all of the YMA based models have over 500 million downloads. Uh but yeah so when we launch new models we want to make sure that people can use the models witβ¦β
βone of things we're actually going to do is in uh next year is we're going to release kis kit 1.0 and that's our commitment to making something more stable more reliable and easier to useβ
βnotion AI is saying I can go ahead and like complete or give you ideas or finish the bullet points for youβ
βI go and check out this Library called scikit-learn it's a very popular machine learning library and I have literally no idea what any of these words mean...β
βlast week we released Claude 3.5 sonnet which on many of the public benchmarks and you know if you just just go and look at social media what people think of it I think it's fair to say of course I'm β¦β
βif you're looking for something like more complicated data analysis really kind of understanding something that you'd have to think about then something like 3.5 Sonet or 3 Opus may be more appropriatβ¦β
βI don't know if you've seen the demos of TL draw um it's this it's this yeah it says interface where it's essentially like a blank canvas and you can use prompting to like and drawing to like generateβ¦β
βfor me the biggest moment was it was chat GPT code interpreter right the mode of chat GPT where it can not just write python code but it can execute that python code and show you the resultsβ
βvoice mode which I love I use that I I I literally I go on an hourong walk with my dog and I'm coding while I'm walking here because I've got airpods in and I can tell talk and voice mode can use codeβ¦β
β# LangMem SDK for agent long-term memoryβ
βthere are other others like versel augment um um um uh cognition uh several several others and so this is happening very fast and you know coders across the industry are using it and the quality of coβ¦β
β# Introducing Command A: Max performance, minimal computeβ
βyou can try Us for free at camel a.com awesome thank youβ
βIf you want access to my AI tool stack - Dovetail, Arize, Linear, Descript, Reforge Build, DeepSky, Relay.app, Magic Patterns, Speechify, and Mobbin - grab Aakashβs bundle .β
βbolt. new just went live I think yesterday with Gemini support cursor has Gemini support now and and is using 2.0 flash hopefully we'll see others like lovable and vzer etc etc have have that support β¦β
βthere is already a shelling Point among among human developers where Tailwind is like the shelling point it seems like now at least according to most peopleβ
βIf you are on one of those paid plans up under the selector here, you should see Thinking 5.4 and Pro 5.4. These are the brand new models that you can use.β
βI love for coding I love cloud code and open increasingly using um OpenAI uh codecs as well.β
βLots of folks thought that GPT 5.5 was already better than Opus 4.8.β
βEmpirically, at a $6.5\times$ smaller parameter budget, QHA generalizes hidden-subset parity of every order $k\le6$ from disjoint inputs, whereas the larger classical attention head collapses past ordβ¦β
βHiViG consistently outperforms existing scalar and verbal critics, improving average success rates over the strongest baseline by 5.8% for Qwen3-VL-32B and 9.0% for Gemini-3-Flashβ
βdemo is available at https://huggingface.co/spaces/Harry-1234/MODF-SIRβ
βOur findings surface significant tensions between the relationality of our participants' queer art practices and the perceived anti-relationality of GenAI development and use. We detail how our particβ¦β
βgsutil mb gs://$PROJECT_NAME-logs-$1β
βwe introduce ATLAS, an annotation tool tailored for long-horizon robotic action segmentation. ATLAS provides time-synchronized visualization of multi-modal robotic dataβ
βThe Strength Pareto Evolutionary Algorithm 2 (SPEA2) is a popular and prominent evolutionary algorithm for solving multi-objective optimisation problems.β
βI love the folks at mrol and I think they're doing really important workβ
βThe Ising framework maps the decoding problem in quantum error correction onto ground-state optimization of a classical Hamiltonian, in which X-Z error correlations enter as cross terms. Under phenomeβ¦β
βwe demonstrate that autoregressive policies can achieve real-time execution by adjusting the tokenization horizon and applying constrained decodingβ
βtoday we brought Gemini 2.0 flash an updated version of it um into production so that developers can actually continue to build with it we announced pricing 10 cents per million input tokens 40 cents β¦β
βNow launching GStack Browseβ
βfrom langchain.chat_models import ChatOpenAIβ
βfrom langchain.embeddings import OpenAIEmbeddingsβ
βfrom langchain.prompts import ChatPromptTemplateβ
βfrom langchain.schema.output_parser import StrOutputParserβ
βfrom langchain.prompts import ChatPromptTemplateβ
βLLM-as-a-judge has become the de facto approach for evaluating LLM outputs.β
βsimilarly like stable diffusion and mid Journey were all like pretty competitive with Dolly 2 and actually probably making more money than DOL doβ
βLast year, Randall launched Alex, an insane AI coaching tool trained solely on his insights. It's effectively an executive coach available 24-7 for only $300 a year. We use it regularly and always forβ¦β
βThree small frozen models (Llama-3.2-1B, Qwen2.5-1.5B, Gemma-2-2B) encode the input into a shared latent space whose aggregate signal is injected into two larger frozen models (Phi-3-mini, Mistral-7B)β¦β
βWe evaluate six CAM techniques: GradCAM, GradCAM++, LayerCAM, EigenCAM, ScoreCAM, and MS GradCAM++ across three CNN architectures (DenseNet201, InceptionV3, ResNet50V2) over thirty training epochs on β¦β
βAmong the CNN models, DenseNet201 has achieved the highest test accuracy of 99%.β
βAnd Rocketpool, the decentralized Ethereum staking protocol, just got a major upgrade, which means now is the best time to stake your ETH for our ETH in return.β
βSo we shipped the first to market transition product or post-quantum, you know, kind of migration product for Bitcoin called Yellow Pages. I think it's still live on Yellow Pages XYZ.β
βI find Bing to often be the most powerful but also to be kind of the scariest and weirdest to use because it has a strong personality that interacts with inter you know your interactions in ways that β¦β
βWe present SPADE, a method for automatically synthesizing data quality assertions that identify bad LLM outputs.β
βFour days ago, we launched ParseBench, the first document OCR benchmark for AI agents.β
βyou can see with music, you can see you know uh Suno and IDO and so forth.β
βcreator of the data set op Source data exploration toolβ
βThere is a a website called the quantum algorithm zoo and it lists of 70 algorithms for which we expect the quantum advantage and then invariably um people yeah like my boy they ask for an area where β¦β
βQ: how do you make diagrams? A: Claude Design - make 5 different SVGs, make a concept and choose one. diagrams, code snippets. use principles of claude design, give design system and references.β
βBy connecting Jira, Confluence, Loom, and Rovo, teams can easily align work to real business goals, plan and track progress together, remove knowledge silos, and access the real magic.β
βPerplexity Comet: https://www.perplexity.ai/cometβ
βExa: https://exa.aiβ
βI I was using hippocratics a a AI system it hasn't yet been released but they showed it to me and it replaces a nurse's systemβ
βperplexity or Google bard or uh chat GPT they all hallucinate so you have to validate everything they tell you that's rule number one if you're going to turn turn in a college essay and try you knowβ
βYou can see this offer in the graphical account viewer: https://www.stellar.org/viewer/#/gdb.β
βMy suggestion, jump over to the tools database, select Matt's picks. These are the tools that I've actually spent a bit of time demoing and testing and vetting myself.β
βOne of the most powerful tools for studying systems like this and many, many others is the Laplace transform. And today, I want to show you exactly how it looks to use this tool to study differential β¦β
βAnd a lot of people kind of -- You know, that was sort of the inspiration, you know, behind Torch, the first Torch, which had the Lua language as the front-end for Caffe2, or for Caffe1, actually, andβ¦β
βWe have whisper that takes an audio and and generates text transcriptsβ
β- [@vitejs/plugin-react-swc](https://github.com/vitejs/vite-plugin-react-swc) uses [SWC](https://swc.rs/) for Fast Refreshβ
βcirc from googleβ
βif the augur and gnosis and all of these other ethereum and blockchain based prediction markets end up launching and it continues to be the case that lots of people don't end up using them then you knβ¦β
βlearn more at arcades.in including videos transcripts and upcoming dates visit conversations withth tyler.comβ
βI remember when when Matt from WordPress I sold my company to WordPress said we have to install slack in deprecate IRCβ