📆 ThursdAI - Jan 30 - DeepSeek vs. Nasdaq, R1 everywhere, Qwen Max & Video, Open Source SUNO, Goose agents & more AI news

Hey folks, Alex here 👋

It’s official—grandmas (and the entire stock market) now know about DeepSeek. If you’ve been living under an AI rock, DeepSeek’s new R1 model just set the world on fire, rattling Wall Street (causing the biggest monetary loss for any company, ever!) and rocketing to #1 on the iOS App Store. This week’s ThursdAI show took us on a deep (pun intended) dive into the dizzying whirlwind of open-source AI breakthroughs, agentic mayhem, and big-company cat-and-mouse announcements. Grab your coffee (or your winter survival kit if you’re in Canada), because in true ThursdAI fashion, we’ve got at least a dozen bombshells to cover—everything from brand-new Mistral to next-gen vision models, new voice synthesis wonders, and big moves from Meta and OpenAI.

We’re also talking “reasoning mania,” as the entire industry scrambles to replicate, dethrone, or ride the coattails of the new open-source champion, R1. So buckle up—because if the last few days are any indication, 2025 is officially the Year of Reasoning (and quite possibly, the Year of Agents, or both!)

Open Source LLMs

DeepSeek R1 discourse Crashes the Stock Market

One-sentence summary: DeepSeek’s R1 “reasoning model” caused a frenzy this week, hitting #1 on the App Store and briefly sending NVIDIA’s stock plummeting in the process ($560B drop, largest monetary loss of any stock, ever)

Ever since DeepSeek R1 launched (our technical coverate last week!), the buzz has been impossible to ignore—everyone from your mom to your local barista has heard the name. The speculation? DeepSeek’s new architecture apparently only cost $5.5 million to train, fueling the notion that high-level AI might be cheaper than Big Tech claims. Suddenly, people wondered if GPU manufacturers like NVIDIA might see shrinking demand, and the stock indeed took a short-lived 17% tumble. On the show, I joked, “My mom knows about DeepSeek—your grandma probably knows about it, too,” underscoring just how mainstream the hype has become.

Not everyone is convinced the cost claims are accurate. Even Dario Amodei of Anthropic weighed in with a blog post arguing that DeepSeek’s success increases the case for stricter AI export controls.

Public Reactions

* Dario Amodei’s blogIn “On DeepSeek and Export Controls,” Amodei argues that DeepSeek’s efficient scaling exemplifies why democratic nations need to maintain a strategic leadership edge—and enforce export controls on advanced AI chips. He sees Chinese breakthroughs as proof that AI competition is global and intense.

* OpenAI Distillation EvidenceOpenAI mentioned it found “distillation traces” of GPT-4 inside R1’s training data. Hypocrisy or fair game? On ThursdAI, the panel mused that “everyone trains on everything,” so perhaps it’s a moot point.

* Microsoft ReactionMicrosoft wasted no time, swiftly adding DeepSeek to Azure—further proof that corporations want to harness R1’s reasoning power, no matter where it originated.

* Government reactedEven officials in the government, David Sacks, US incoming AI & Crypto czar, discussed the fact that DeepSeek did "distillation" using the term somewhat incorrectly, and presidet Trump was asked about it.

* API OutagesDeepSeek’s own API has gone in and out this week, apparently hammered by demand (and possibly DDoS attacks). Meanwhile, GPU clouds like Groq are showing up to accelerate R1 at 300 tokens/second, for those who must have it right now.

We've seen so many bad takes on the topic, from seething cope takes, to just gross misunderstandings from gov officials confusing the ios App with the OSS models, folks throwing conspiracy theories into the mix, claiming that $5.5M sum was a PsyOp. The fact of the matter is, DeepSeek R1 is an incredible model, and is now powering (just a week later), multiple products (more on this below) and experiences already, while pushing everyone else to compete (and give us reasoning models!)

Open Thoughts Reasoning Dataset

One-sentence summary: A community-led effort, “Open Thoughts,” released a new large-scale dataset (OpenThoughts-114k) of chain-of-thought reasoning data, fueling the open-source drive toward better reasoning models.

Worried about having enough labeled “thinking” steps to train your own reasoner? Fear not. The OpenThoughts-114k dataset aggregates chain-of-thought prompts and responses—114,000 of them—for building or fine-tuning reasoning LLMs. It’s now on Hugging Face for your experimentation pleasure. The ThursdAI panel pointed out how crucial these large, openly available reasoning datasets are. As Wolfram put it, “We can’t rely on the big labs alone. More open data means more replicable breakouts like DeepSeek R1.”

Mistral Small 2501 (24B)

One-sentence summary: Mistral AI returns to the open-source spotlight with a 24B model that fits on a single 4090, scoring over 81% on MMLU while under Apache 2.0.

Long rumored to be “going more closed,” Mistral AI re-emerged this week with Mistral-Small-24B-Instruct-2501—an Apache 2.0 licensed LLM that runs easily on a 32GB VRAM GPU. That 81% MMLU accuracy is no joke, putting it well above many 30B–70B competitor models. It was described as “the perfect size for local inference and a real sweet spot,” noting that for many tasks, 24B is “just big enough but not painfully heavy.” Mistral also finally started comparing themselves to Qwen 2.5 in official benchmarks—a big shift from their earlier reluctance, which we applaud!

Berkeley TinyZero & RAGEN (R1 Replications)

One-sentence summary: Two separate projects (TinyZero and RAGEN) replicated DeepSeek R1-zero’s reinforcement learning approach, showing you can get “aha” reasoning moments with minimal compute.

If you were wondering whether R1 is replicable: yes, it is. Berkeley’s TinyZero claims to have reproduced the core R1-zero behaviors for $30 using a small 3B model. Meanwhile, the RAGEN project aims to unify RL + LLM + Agents with a minimal codebase. While neither replication is at R1-level performance, they demonstrate how quickly the open-source community pounces on new methods. “We’re now seeing those same ‘reasoning sparks’ in smaller reproductions,” said Nisten. “That’s huge.”

Agents

Codename Goose by Blocks (X, Github)

One-sentence summary: Jack Dorsey’s company Blocks released Goose, an open-source local agent framework letting you run keyboard automation on your machine.

Ever wanted your AI to press keys and move your mouse in real time? Goose does exactly that with AppleScript, memory extensions, and a fresh approach to “local autonomy.” On the show, I tried Goose, but found it occasionally “went rogue, trying to delete my WhatsApp chats.” Security concerns aside, Goose is significant: it’s an open-source playground for agent-building. The plugin system includes integration with Git, Figma, a knowledge graph, and more. If nothing else, Goose underscores how hot “agentic” frameworks are in 2025.

OpenAI’s Operator: One-Week-In

It’s been a week since Operator went live for Pro-tier ChatGPT users. “It’s the first agent that can run for multiple minutes without bugging me every single second,”. Yet it’s still far from perfect—captchas, login blocks, and repeated confirmations hamper tasks. The potential, though, is enormous: “I asked Operator to gather my X.com bookmarks and generate a summary. It actually tried,” I shared, “but it got stuck on three links and needed constant nudges.” Simon Willison added that it’s “a neat tech demo” but not quite a productivity boon yet. Next steps? Possibly letting the brand-new reasoning models (like O1 Pro Reasoning) do the chain-of-thought under the hood.

I also got tired of opening hundreds of tabs for operator, so I wrapped it in a macOS native app, that has native notifications and the ability to launch Operator tasks via a Raycast extension, if you're interested, you can find it on my Github

Browser-use / Computer-use Alternatives

In addition to Goose, the ThursdAI panel mentioned browser-use on GitHub, plus numerous code interpreters. So far, none blow minds in reliability. But 2025 is …