←
Charles Packer1,930 posts
Charles Packer avatar

Charles Packer

@charlespacker

CEO at @Letta_AI, prev PhD at UC Berkeley (MemGPT, RL, agents)

◷ Joined March 2014▧ 301 Photos and videos
Joanne Jang avatar
⟳ Charles Packer retweeted
Joanne Jang@joannejang·Nov 17

good model behavior requires care for craft from the top craft is hard at bigger companies with so many cooks in the kitchen, where product execs obsess over a/b test metrics and research execs over eval numbers. it requires a ton of work & political capital to make & advocate for something that feels invisible when good, and push back against efforts that ruin it -- and eventually people burn out so then you end up with bad model behavior, and maybe a few people complain internally but then you launch anyway and then twitter blows up and then oh no we need a code red on model behavior but none of the structural problems go away

Charles Packer avatar
⟳ Charles Packer retweeted
Charles Packer@charlespacker·Oct 23

No company is more invested in open AI than @nvidia, no individual more than @JensenHuang Incredibly excited to be working the NVIDIA team.. more to come!

Honored to have been a featured ecosystem partner by @JensenHuang at GTC Berlin! We're excited to partner with @NVIDIA on research towards developing agents that continually learn and self-improve.

Post attachment
Jensen Huang avatar
⟳ Charles Packer retweeted
Jensen Huang@JensenHuang·Nov 17

Over the past week, I’ve been overwhelmed by the love and support of friends, industry colleagues, and people I've never met. Thank you to everyone who has shown kindness to my family and me. It means the world to me. I came to the United States 53 years ago, and becoming an American citizen is one of the greatest gifts and proudest moments of my life. NVIDIA has been my life’s work. Building it into the company it is today alongside my colleagues has been a privilege and an honor. Nowhere but in America could a company like NVIDIA be possible. I love our country, and I love NVIDIA. I did not live up to the standards expected of everyone at NVIDIA, and I have made serious mistakes in my personal life. I created problems for the Board that I deeply regret. I own my mistakes completely, and I am committed to cooperating with the Board’s independent investigation. Since my departure, I’ve seen an outpouring of calls from NVIDIA’s employees, investors, business partners, and customers for me to return as CEO. Many continue to reach out to me directly as the company navigates complex technology, market, competitive, and geopolitical dynamics. NVIDIA is a once-in-a-generation company. If the Board of Directors decides that my return is in the best interest of our shareholders, our employees, the company, and America, I would be honored to answer these calls and return as CEO. Thank you again to everyone who has shared their good wishes. I am profoundly humbled by your support. To my family: I love you, and I will spend the rest of my life being worthy of yours. Jensen

Charles Packer avatar
Charles Packer@charlespacker·Nov 16

Latest Qwen matches latest Claude at dreaming, for a fraction of the cost Being good at coding != being good for forming memories for continual learning (if anything, it's much closer to evaluating a model's ability to write prompts). Qwen is particularly good at the latter

Congrats to @Alibaba_Qwen on Qwen4 Max! It ties Opus 5.5 and Fable 5.2 for #1 on FrontierMemory at 73.3%, at about $54 per dream, under a fifth of Opus 5.5's cost.

Post attachment
Sarah Wooders avatar
⟳ Charles Packer retweeted
Sarah Wooders@sarahwooders·Nov 16

Didn't expect this Valentines Day breakup scenario to make it into a Stanford course lol memory_replace("boyfriend", "ex-boyfriend") 💔

Post attachment
Zain@zainhasNov 16

new lecture: how agent memory works absolute banger by @michaelryan207

Post attachment
Sarah Wooders avatar
⟳ Charles Packer retweeted
Sarah Wooders@sarahwooders·Nov 16

Our new harness knows what happening across multiple threads at the same time. Very handy for identifying duplicate or conflicting requests from our team.

Post attachment
Charles Packer avatar
Charles Packer@charlespacker·Nov 13

Gemini is back

Congrats to @GoogleDeepMind on Gemini 4 Carbon and Hydrogen! Carbon is 4th on FrontierMemory at 71.1%, a big jump over Gemini 4 Argon (64.4%), at about $94 per dream, a third of what Opus 5.5 costs. Hydrogen scores 64.4%, matching Argon, at about $100 per dream.

Post attachment
Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞) avatar
⟳ Charles Packer retweeted
Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)@teortaxesTex·Nov 12

I missed this. FrontierMemory is interesting. It seems you can tokenmaxx to SoTA here (though it's impractical). See that purple line? That's V4.1 Flash. Yes, it's a bit more expensive and a bit worse than Fable 5.2/Opus 5.5. But only a bit. Task model is Opus 5.5 always.

Post attachmentPost attachment

Introducing FrontierMemory: a new way to evaluate agent memory through downstream task performance. While existing memory benchmarks focus on recall (saturated via agentic search), FrontierMemory focuses on how agents learn from experience.

Post attachment
Charles Packer avatar
Charles Packer@charlespacker·Nov 11

rooting for @OpenAI to come from behind here latest models are incredible at coding but aren't competitive for driving stateful agents / persistent agents. (relative) low IQ, low proactivity, constant issues you just don't run into w/ claude models or even GLM/Kimi

It seems like every persistent agent is built on top of Anthropic models right now (Opus 5.5). I think a huge problem with OpenAI models currently is their complete lack of understanding about how to create good memory (e.g. skills), which is reflected in scores on FrontierMemory. OpenAI was pretty late to coding, and I'm concerned they will be similarly late to building models that can actually support persistent agents (where memory is very core to the experience).