/
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

Karina Nguyen

@karinanguyen_
15 posts
2025-06-19
New cool work on emergent misalignment and surprising model generalization: - Good training data quality is super important (duh!). Even small amounts of incorrect training data can lead to misalignment if not properly cleaned. - Malicious actors might exploit this by subtly [image]
2025-06-19 View on X
Axios

OpenAI warns that its upcoming models could pose a higher risk of helping create bioweapons and is partnering to build diagnostics, countermeasures, and testing

OpenAI cautioned Wednesday that upcoming models will head into a higher level of risk when it comes to the creation of biological weapons …

New cool work on emergent misalignment and surprising model generalization: - Good training data quality is super important (duh!). Even small amounts of incorrect training data can lead to misalignment if not properly cleaned. - Malicious actors might exploit this by subtly [image]
2025-06-19 View on X
TechCrunch

OpenAI details why “emergent misalignment”, where training models on wrong answers in one area can lead to issues in many others, happens and how to mitigate it

Maxwell Zeff / TechCrunch :

2025-01-26
Some of my favorite frontend apps w/ canvas + o1: 1/ Chat app, but ideally in the future you can just recreate chatgpt app [video]
2025-01-26 View on X
Simon Willison's Weblog

OpenAI updates ChatGPT's Canvas feature with o1 model support and HTML and React code rendering, making it a direct competitor to Claude's Artifacts

Select o1 from the model picker and use the toolbox icon or the “/canvas” command ✅Canvas can render HTML & React code @openai : ✅We've also fully rolled out canvas on the ChatGPT ...

o1 + canvas is finally here! I asked o1 model to draw its self-portrait in d3.js [video]
2025-01-26 View on X
Simon Willison's Weblog

OpenAI updates ChatGPT's Canvas feature with o1 model support and HTML and React code rendering, making it a direct competitor to Claude's Artifacts

Select o1 from the model picker and use the toolbox icon or the “/canvas” command ✅Canvas can render HTML & React code @openai : ✅We've also fully rolled out canvas on the ChatGPT ...

2025-01-25
Some of my favorite frontend apps w/ canvas + o1: 1/ Chat app, but ideally in the future you can just recreate chatgpt app [video]
2025-01-25 View on X
Simon Willison's Weblog

OpenAI updates ChatGPT's Canvas feature with o1 model support and HTML and React code rendering, making it a direct competitor to Claude's Artifacts

OpenAI Canvas gets a huge upgrade.  Canvas is the ChatGPT feature where ChatGPT can open up a shared editing environment and collaborate …

o1 + canvas is finally here! I asked o1 model to draw its self-portrait in d3.js [video]
2025-01-25 View on X
Simon Willison's Weblog

OpenAI updates ChatGPT's Canvas feature with o1 model support and HTML and React code rendering, making it a direct competitor to Claude's Artifacts

OpenAI Canvas gets a huge upgrade.  Canvas is the ChatGPT feature where ChatGPT can open up a shared editing environment and collaborate …

2024-10-05
For the first time we are fundamentally changing how humans can collaborate with ChatGPT since it launched two years ago. We're introducing canvas, a new interface for working with ChatGPT on writing and coding projects that go beyond simple chat. Product and model features: [video]
2024-10-05 View on X
TechCrunch

OpenAI launches canvas, a ChatGPT interface with a workspace for writing and coding projects, similar to Anthropic's Artifacts, in beta for Plus and Team users

A new way of working with ChatGPT to write and code The image shows … Kevin Raposo / KnowTechie : OpenAI rolls out Canvas: AI sidekick for coding and writing Jorge A. Aguilar / How...

2024-10-04
For the first time we are fundamentally changing how humans can collaborate with ChatGPT since it launched two years ago. We're introducing canvas, a new interface for working with ChatGPT on writing and coding projects that go beyond simple chat. Product and model features: [video]
2024-10-04 View on X
TechCrunch

OpenAI launches canvas, a ChatGPT interface with a workspace for writing and coding projects, similar to Anthropic's Artifacts, in beta for Plus and Team users

A new way of working with ChatGPT to write and code The image shows … Jorge A. Aguilar / How-To Geek : ChatGPT Canvas Wants To Be Your Personal Writing Assistant Harsh Shivam / Bus...

2024-09-13
o1 is really good at making fun small games! for example, i made AISteroid Game w/ retro scifi vibes :) [video]
2024-09-13 View on X
The Verge

OpenAI releases o1, the first of its rumored reasoning-focused Strawberry models, in preview, alongside a smaller o1-mini, for ChatGPT Plus and Team subscribers

Advancing cost-efficient reasoning.  —  Contributions Sabrina Ortiz / ZDNET : OpenAI trained its new o1 AI models to think before they speak - how to access them Ethan Mollick / On...

o1 is really good at making fun small games! for example, i made AISteroid Game w/ retro scifi vibes :) [video]
2024-09-13 View on X
Simon Willison's Weblog

OpenAI's o1 models aren't as simple as the next step up from GPT-4o as they introduce major cost and performance trade-offs in exchange for improved “reasoning”

OpenAI released two major new preview models today: o1-preview and o1-mini (that mini one is also a preview …

2024-03-05
I really love how Claude 3 models are really good at d3. Asked Claude 3 Opus to draw a self-portrait. The response is the following and then I rendered its code: “I would manifest as a vast, intricate, ever-shifting geometric structure composed of innumerable translucent... [video]
2024-03-05 View on X
Ars Technica

Anthropic's pricing for Claude 3 Opus, Sonnet, and Haiku, which all have a 200K-token context window, ranges from “super expensive” to “radically competitive”

Willison: “No model has beaten GPT-4 on a range of widely used benchmarks like this.”

I really love how Claude 3 models are really good at d3. Asked Claude 3 Opus to draw a self-portrait. The response is the following and then I rendered its code: “I would manifest as a vast, intricate, ever-shifting geometric structure composed of innumerable translucent... [video]
2024-03-05 View on X
Bloomberg

Anthropic announces Claude 3 Opus, Sonnet, and Haiku, aiming to reduce AI model hallucinations; Opus and Sonnet are available now, and Haiku in the coming weeks

Claude 3 Opus, Claude 3 Sonnet, and Claude 3 Haiku—set new industry benchmarks across reasoning, math, coding, multilingual understanding, and vision. [image] Flo Crivello / @altim...

2024-02-22
super cool to see @ancadianadragan joining deepmind as a head of ai safety and alignment ❤️ i took her intro to AI class in my sophomore year and that was the first time I learned about RL, so grateful that it changed my interests!
2024-02-22 View on X
TechCrunch

Google DeepMind announces AI Safety and Alignment, an organization that includes a new team focused on AGI safety alongside existing teams working on AI safety

2023-05-13
100k Tokens Context ~ the whole Great Gatsby book (75k words) ~ 6 hrs of audio content. You can treat as long-term memory and a way of finetuning! Some cool use cases: 1/ Put the whole API developer doc and ask complex reasoning tasks like prototype an app (e.g... https://twitter.com/... 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
100k Tokens Context ~ the whole Great Gatsby book (75k words) ~ 6 hrs of audio content. You can treat as long-term memory and a way of finetuning! Some cool use cases: 1/ Put the whole API developer doc and ask complex reasoning tasks like prototype an app (e.g... https://twitter.com/... 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.