/
Navigation
C
Chronicles
Browse all articles
C
E
Explore
Semantic exploration
E
R
Research
Entity momentum
R
N
Nexus
Correlations & relationships
N
~
Story Arc
Topic evolution
S
Drift Map
Semantic trajectory animation
D
P
Posts
Analysis & commentary
P
Browse
@
Entities
Companies, people, products, technologies
Domains
Browse by publication source
Handles
Browse by social media handle
Detection
?
Concept Search
Semantic similarity search
!
High Impact Stories
Top coverage by position
+
Sentiment Analysis
Positive/negative coverage
*
Anomaly Detection
Unusual coverage patterns
Analysis
vs
Rivalry Report
Compare two entities head-to-head
/\
Semantic Pivots
Narrative discontinuities
!!
Crisis Response
Event recovery patterns
Connected
Nav: C E R N
Search: /
Command: ⌘K
Embeddings: large
VOICE ARCHIVE

Andrew Gao

@itsandrewgao
18 posts
2025-10-31
don't write this off as “fast, non-frontier-lab model == dumb & not worth my time” it's smarter than the SOTA models were this summer and also way faster (more chances to iterate/fix in same time, less waiting) 1 pt of reference: SWE-1.5 > GPT-5 (high) on SWE-Bench Pro! [image]
2025-10-31 View on X
Cognition

Cognition releases SWE-1.5, a new coding model in Windsurf, saying it partnered with Cerebras to serve SWE-1.5 at speeds up to 13x faster than Claude Sonnet 4.5

lots of new paradigms/UX to figure out. @cognition : Today we're releasing SWE-1.5, our fast agent model. It achieves near-SOTA coding performance while setting a new standard for ...

keep in mind: the August snapshot of Opus 4.1 scored 22.71% on SWE-Bench Pro. SWE-1.5 (13x faster than sonnet 4.5) scores ~2x higher than the SOTA code model from only 2 months ago. made possible by owning the stack: model, inference, & agent harness! agent lab era is here! [image]
2025-10-31 View on X
Cognition

Cognition releases SWE-1.5, a new coding model in Windsurf, saying it partnered with Cerebras to serve SWE-1.5 at speeds up to 13x faster than Claude Sonnet 4.5

lots of new paradigms/UX to figure out. @cognition : Today we're releasing SWE-1.5, our fast agent model. It achieves near-SOTA coding performance while setting a new standard for ...

2025-05-22
we've made an official DeepWiki remote MCP server! [image]
2025-05-22 View on X
VentureBeat

OpenAI updates its Responses API for building agentic applications to include remote MCP server support, image generation and Code Interpreter tools, and more

OpenAI is rolling out a set of significant updates to its newish Responses API, aiming to make it easier for developers and enterprises …

2024-05-15
Never bet against @Google! They just dropped a Sora competitor. 1080p, over a minute long vids, and impeccable quality All the #Veo examples so far 🧵👇 [video]
2024-05-15 View on X
CNET

Google unveils Music AI Sandbox, a suite of music tools that include the ability to create loops via text, without disclosing a public launch date

Exciting news from Google at #IO2024 🎉 We introduced #Veo and #Imagen3 … See also Mediagazer

Never bet against @Google! They just dropped a Sora competitor. 1080p, over a minute long vids, and impeccable quality All the #Veo examples so far 🧵👇 [video]
2024-05-15 View on X
Engadget

Google debuts Veo, a text-to-video model that outputs at 1080p resolution, and Imagen 3, its highest quality image generation model with improved text rendering

2024-05-14
no offense to the lovely team at openai but i could really have cared less about the GPT store. This is what we wanna see! [image]
2024-05-14 View on X
TechCrunch

OpenAI unveils GPT-4o, a new flagship generative AI model that is faster and natively multimodal, rolling out for free to all ChatGPT users in the coming weeks

There are two things from our announcement today I wanted to highlight. OpenAI : Hello GPT-4o  —  We're announcing GPT-4o, our new flagship model that can reason across audio, visi...

no offense to the lovely team at openai but i could really have cared less about the GPT store. This is what we wanna see! [image]
2024-05-14 View on X
The Verge

OpenAI opens its GPT Store to all users for free, including the ability to create custom GPTs; OpenAI debuted GPT Store for paid users on January 10

2024-05-01
uh.... gpt2-chatbot just solved an International Math Olympiad (IMO) problem in one-shot the IMO is insanely hard. only the FOUR best math students in the USA get to compete prompt + its thoughts 🧵 [image]
2024-05-01 View on X
Gizmodo

A look at gpt2-chatbot, a mysterious AI chatbot that became available on LLM benchmarking site LMSYS Org and is believed to have similar capabilities as GPT-4

An advanced AI model with unknown origins turned heads this week as online communities unpacked a cryptic tweet from Sam Altman.

Why are people mad that Rabbit is using Android OS? Feels overkill to build an entire os from scratch.
2024-05-01 View on X
Android Authority

An APK review suggests Rabbit's R1 runs Android, and its entire UI is powered by a single Android app; founder Jesse Lyu says “rabbit r1 is not an Android app”

The Rabbit R1 is probably running Android and is certainly powered by an Android app under the hood.  —  •  —  TL;DR

2024-04-11
Mistral's surprise model is 8x22B which is 176billion params, like GPT3.5 Very excited about this given that Mixtral, at 56B, is already close/surpassing 3.5 So new mistral will be even better, perhaps approaching gpt4
2024-04-11 View on X
VentureBeat

Mistral AI launches Mixtral 8x22B, its latest sparse mixture-of-experts model, after releasing Mixtral 8x7B in December 2023

As Google unleashed a barrage of artificial intelligence announcements at its Cloud Next conference, Mistral AI decided to jump into action with the launch …

2024-03-18
grok is 300GB for reference LLAMA-65b is 120 GB How big is Grok?? (ofc some of the 300 is auxiliary files) [image]
2024-03-18 View on X
xAI

xAI open sources the base model weights and network architecture of Grok-1, a 314B parameter Mixture-of-Experts model trained in October 2023, under Apache 2.0

We are releasing the base model weights and network architecture of Grok-1, our large language model.

here's your DEEP DIVE into @grok's architecture! I just went through the https://model.py/, for this 314B open source behemoth with *no strings attached*. 👇🧵 [image]
2024-03-18 View on X
xAI

xAI open sources the base model weights and network architecture of Grok-1, a 314B parameter Mixture-of-Experts model trained in October 2023, under Apache 2.0

We are releasing the base model weights and network architecture of Grok-1, our large language model.

2023-11-17
it's over. we can now generate anime with AI. the level of cohesiveness from @meta emu video is insane. [video]
2023-11-17 View on X
SiliconANGLE

Meta unveils new AI tools to edit images and generate videos from text instructions, based on Emu, its model for image generation announced in September 2023

Artificial intelligence researchers from Meta Platforms Inc. said today they have made significant advances in AI-powered image and video generation.

2023-05-13
Anthropic's Claude AI blows past #GPT4 with a new 100,000 token context window (75,000 words). Enough to fit Harry Potter Book 1! Unsure how quality degrades (if it does) over length of convo. But if context windows can infinitely scale, embeddings may be in a little trouble.... https://twitter.com/...
2023-05-13 View on X
TechCrunch

Anthropic expands Claude's context window from 9K to 100K tokens, or ~75K words it can digest and analyze; OpenAI's GPT-4 has a context window of ~32K tokens

2023-05-12
Anthropic's Claude AI blows past #GPT4 with a new 100,000 token context window (75,000 words). Enough to fit Harry Potter Book 1! Unsure how quality degrades (if it does) over length of convo. But if context windows can infinitely scale, embeddings may be in a little trouble.... https://twitter.com/...
2023-05-12 View on X
TechCrunch

Anthropic expands Claude's context window from 9K to 100K tokens, or ~75K words it can digest and analyze; OpenAI's GPT-4 has a context window of ~32K tokens

Historically and even today, poor memory has been an impediment to the usefulness of text-generating AI.

2023-02-08
im very bullish on @Meta. Here's why Video game dev and VR content creation is very time intensive because of the 3D/4D. nature. It takes forever to make 3D assets and graphics. @MetaAI's text to 3D model basically 10x's the speed at which developers can create VR content.
2023-02-08 View on X
New York Times

Long an AI leader, Meta has been cautious with chatbots, closing Galactica in November over toxic content and misinformation, letting OpenAI take center stage

New York Times : Tweets: @cademetz , @rao2z , @pstasiatech , and @nytimestech Tweets: Cade Metz / @cademetz : Two weeks before the arrival of ChatGPT, Meta released its own bot. T...

Never thought I'd be downloading Bing in 2023... @microsoft is back! https://twitter.com/...
2023-02-08 View on X
The Verge

Microsoft unveils an updated Bing search engine and Edge browser powered by a next-gen custom OpenAI LLM, calling the duo an “AI-powered co-pilot for the web”

Microsoft has announced a new version of its search engine Bing, powered by an upgraded version of the same AI technology that underpins chatbot ChatGPT.

2023-01-28
Google's Text to Music model is really cool. They're kind of a “sleeping giant” it seems and have been making big moves in relative silence. Imagen, PaLM, now this. https://google-research.github.io/ ...
2023-01-28 View on X
TechCrunch

Google researchers detail AI model MusicLM, which can generate high-fidelity music in any genre from text and was trained on a dataset of 280K hours of music

An impressive new AI system from Google can generate music in any genre given a text description.  But the company, fearing the risks, has no immediate plans to release it.