Brian Zhan (@brianzhan1) started following @nomic_ai on Oct 6, 2024

Exciting to see the state-of-the-art in text embedding being built over Nomic open research! Open source results in massive impact

Explore the lives and legacies of over 1.4 million individuals from English Wikipedia’s biographical articles in our Atlas map here: https://t.co/Rq5cBLwxrn

#ectel2024

Attending #ec-tel in Austria. Heard there is heavy rain in Vienna and got help at this Sushi place in London airport by a kind lady! https://t.co/eQPEno7fP6

Though useful and powerful, an LLM today trained on virtually all the text on the internet is still far from human-level. Do you think there exists *any* architecture and algorithm that when trained on this same corpus would surpass humans? Or is the data simply inadequate?

RT @MLStreetTalk: 🧠 Is the future of AI open-ended? Latest show with @_rockt (who achieved two ICML best paper awards! 📷https://t.co/tfYVK…

audible should play a “previously on” segment when you go back to a book you haven’t listened to in a while

new wallpaper, courtesy of this morning's sunrise https://t.co/JHp6MpBRSp

Switched most of my IRL meetings to walking ones lately and it’s clearly the best meeting format. All meetings should be walks. Need to present? Use your phone. Need to whiteboard? Pace around the conference room. 3 hour board meeting? Saddle up for a 15 mile trek!

Emerging platforms that support the creation of physical-world applications are mutually dependent. e.g. XR headsets as visual interface + data collection; sensors/wearables as read input; robotic manipulator as write output. These shouldn't be thought of as discrete pieces.

RT @VentureBeat: Mabl launches an AI-powered low-code mobile app testing service https://t.co/9elzBO86pH https://t.co/Bgt0ppyokn

RT @verycoolgreg: @mablhq Unlike traditional testing frameworks, mabl's use of machine intelligence ensures that expanded testing doesn't…

RT @gabepereyra: @harvey__ai's tax foundation built with @PwC is preferred by tax experts 91% of the time. Huge shoutout to @calvincongela…

RT @anaganath: Successful 3 table poker night co-hosted with @sanjaypartly ! DM either of us if you’d like to join the next one 😊 https://t…

RT @zhegan4: 💡CLIP is the default choice for most multimodal LLM research. But, we know CLIP is not perfect. It is good at high-level seman…

RT @RohitGandikota: What should be the goal of unlearning in language models? In our new preprint we look at this question carefully and pr…

RT @arpit_bhayani: Happy Engineer's Day everyone! it is a great day to remind ourselves that we are fortunate to be part of one of the mos…

Now we would be waiting for @GoogleAI and @AnthropicAI to release their models. But, I guess the edge is there with @OpenAI and they will always be miles ahead of both of them. I bet that @xai could be a game changer at this point. @elonmusk has always been a underdog.

Ok, so PG's founder mode essay was pretty bang on. Probably should apply that same advice to YC itself though, no?

@mymagicpin @ola_supports @Olacabs @bhash @bhash dawg what kind of fraud you running wtf

RT @JuntaoRen: Excited to share a new framework to design provably efficient inverse reinforcement learning algorithms, to appear at #ICML2…

RT @DanelDayan: Databricks disclosed that they are expecting to cross $2.4B in revenue run-rate growing ~65% Y/Y by H1’25 (Jan FYE). This w…

Amazing fun putting this talk together on some of the nuts and bolts of how we built Genie, thanks again to @OpenAI for the opportunity!

RT @CosineAI: o1 is here! @AlistairPullen sat down with @shyamalanadkat at the @OpenAI office to discuss how our unique techniques have unl…

Part of what made us excited about @VoyageAI was the customer love and recognition from AI platform leaders like @vivek7ue, SVP of Engineering of Snowflake, that Voyage models are the best for RAG. We're incredibly excited to partner on the Series A alongside @SnowflakeDB

Part of what made us excited about @VoyageAI was the customer love and recognition from platform leaders like @vivek7ue, SVP of Engineering of Snowflake, that Voyage models are the best. We're incredibly excited to partner on the Series A alongside @SnowflakeDB and others

RT @SnowflakeDB: “Voyage AI ‘s models are far higher quality for the really hard use cases.” - @vivek7ue Read more in @VentureBeat on how…

RT @ajay_frontiers: Agree! A friend asked my investment thesis on @VoyageAI . Thesis is simple: Business world has chosen RAG (over fine tu…

RT @Gradio: Fuyu-Heavy from @AdeptAILabs looks very promising -- this demo shows that the model excels at multimodal reasoning and has a br…

RT @AdeptAILabs: Introducing Fuyu-Heavy, our new multimodal model. Fuyu-Heavy is the world’s third-most-capable multimodal model, behind on…