Member of Technical Staff at @openai. Ex @Harvard @Cambridge_Uni. Views my own
San Fransisco, USA
Joined May 2014
- Tweets63
- Following283
- Followers359
- Likes281
Max Murphy retweeted
We have paused some frontier RL training to ensure that we can meet the appropriate alignment, security and monitoring standards for the new level of capabilities in front of us. Model progress is now extremely rapid, and we always said we would take action if we felt that model capabilities were outstripping the pace of safety and alignment.
We care very deeply about AI safety. We believe the entire field will have to coordinate on shared safety standards, but will act unilaterally in the meantime.
We expect confidence in safety to increasingly set the pace of AI progress. We are optimistic about the alignment work we are doing, and we remain committed to making frontier capabilities widely available.
openai.com/index/pacing-mode…
Max Murphy retweeted
Work at OpenAI is being transformed by agents, in every department.
Across our entire company, people are using Codex to do work that is more complex, longer-running, and increasingly cross-functional.
Our internal usage offers an early look at how agentic tools may reshape work as they become more capable and broadly available.
Max Murphy retweeted
I am Tibo and I have an incredible team. Codex would not exist without them and they cooked.
Enjoy the new Codex app, access through your free/go ChatGPT plan and 2X rate limits on other plans. Can't wait to hear what you do with it.
openai.com/codex/
Worked on ChatGPT engineering this year helping to build Pulse
It’s designed to help you achieve your goals by proactively doing work for you
I’ve been using to plan holidays, curate newsletters & get updates on rugby results
Was a lot of fun!!
Today we are launching my favorite feature of ChatGPT so far, called Pulse. It is initially available to Pro subscribers.
Pulse works for you overnight, and keeps thinking about your interests, your connected data, your recent chats, and more. Every morning, you get a custom-generated set of stuff you might be interested in.
It performs super well if you tell ChatGPT more about what's important to you. In regular chat, you could mention “I’d like to go visit Bora Bora someday” or “My kid is 6 months old and I’m interested in developmental milestones” and in the future you might get useful updates.
Think of treating ChatGPT like a super-competent personal assistant: sometimes you ask for things you need in the moment, but if you share general preferences, it will do a good job for you proactively.
This also points to what I believe is the future of ChatGPT: a shift from being all reactive to being significantly proactive, and extremely personalized.
This is an early look, and right now only available to Pro subscribers. We will work hard to improve the quality over time and to find a way to bring it to Plus subscribers too.
Huge congrats to @ChristinaHartW, @_samirism, and the team for building this.
Max Murphy retweeted
We’re sharing how we think about safety and alignment.
We see AGI development as a continuous path rather than a sudden leap. Instead of preparing for a single pivotal moment, we're iteratively deploying and learning from today's models to make future AI safer and more beneficial.
Our approach is guided by embracing uncertainty, defense in depth, methods that scale, human control, and community efforts to ensure that AGI benefits all of humanity.
Max Murphy retweeted
I am not an AI safety researcher so I have no skin in the game but several forces point to the value of and demand for safety work shooting up like crazy in the next year or two:
1. With each new generation of models, capabilities grow, leading to increased surface area for safety risks
2. As we move from chatbots that just chat to agents that take actions (i.e., read-only to read-write), error margins will become smaller as the maximum potential damage of one bad action becomes larger
3. Companies are deploying AI in more products than ever before
Product launches will be bottlenecked on safety work and it will be a race to build competent and efficient safety teams. Good news is that safety researchers can leverage the same new paradigms that unlock new capabilities (e.g., CoT models) to also make safety work faster, more accurate, and more flexible
Can’t get my head round how the people I work with built this. Sora uses AI to create videos. Here are some of the coolest ones I’ve made…
A bear playing the piano in the ocean
Can AI predict war?
I’ve spent a lot of time thinking about this, and have finally got a paper out on it
Very grateful to @TFletcher for his support & for not making me feel crazy when I raised the idea with him many years ago! And to @kaylanhua & Ezra!
cambridge.org/core/journals/…
Not me thinking the ‘congrats on o1’ signs at work were a pat on the back for my new visa
(Turns out they’re for the release of our most capable and aligned models to date…)
Such an exciting day! Try it you might like it…
We're releasing a preview of OpenAI o1—a new series of AI models designed to spend more time thinking before they respond.
These models can reason through complex tasks and solve harder problems than previous models in science, coding, and math. openai.com/index/introducing…
Max Murphy retweeted
Excited to share that the paper Ezra, @maxmurphy100, and I worked on has been published in the Data and Policy journal! In it, we advocate for the use of ML as a tool in geopolitical decision-making and explore the current state of the space. (1/2)
cambridge.org/core/journals/…
