Embed Notice

HTML Code

<blockquote style="position: relative; padding-left: 55px;"><section><a href="https://theforkiverse.com/users/Casey/statuses/116377632736324779">Casey Newton (casey@theforkiverse.com)'s status on Tuesday, 14-Apr-2026 11:02:29 JST</a><a href="https://theforkiverse.com/@Casey" title="casey@theforkiverse.com"><img src="https://gnusocial.jp/avatar/377238-48-20260109234416.webp" width="48" height="48" alt="Casey Newton" style="position: absolute; left: 0; top: 0;">Casey Newton</a></section><article><p>Meta may be back in the AI race — but the race is speeding up, and its rivals' advantages may be compounding. I wrote about Muse Spark: <a href="https://www.platformer.news/meta-muse-spark-ai-race/" rel="nofollow">https://www.platformer.news/meta-muse-spark-ai-race/</a></p></article><footer><a rel="bookmark" href="https://gnusocial.jp/conversation/6319737#notice-12447183">In conversation</a><time datetime="2026-04-14T11:02:29+09:00" title="Tuesday, 14-Apr-2026 11:02:29 JST">about 2 months ago</time> <span>from <span><a href="https://theforkiverse.com/@Casey/116377632736324779" rel="external" title="Sent from theforkiverse.com via ActivityPub">theforkiverse.com</a></span></span><a href="https://theforkiverse.com/@Casey/116377632736324779">permalink</a><h4>Attachments</h4><ol><li><label><a rel="external" href="https://gnusocial.jp/attachment/5991585">But the frontier has not been waiting for Meta to catch up with it. The day before Meta’s launch, Anthropic announced Claude Mythos Preview, a model so capable at discovering software vulnerabilities that it would not release it to the general public. The flip side of that is that the model is extremely capable at agentic coding — it scored 93.9% on SWE-bench Verified, compared to 77.4% for Muse Spark — and Anthropic is already presumably using it heavily as it builds its next set of models and products.OpenAI will reportedly release a similar model soon, and will also restrict its use. (That model is distinct from another next-generation model, codenamed “Spud,” said to be arriving soon.) Overnight, the true frontier in AI development became a model that’s too good to ship. And because those labs can use those models in development, the gap between them and everyone else may be compounding.</a></label><br><a href="https://cdn.masto.host/theforkiversecom/media_attachments/files/116/377/632/460/311/715/original/8eed0056927628db.png" rel="external">https://cdn.masto.host/theforkiversecom/media_attachments/files/116/377/632/460/311/715/original/8eed0056927628db.png</a></li><li><article><header><div>Domain not in remote thumbnail source whitelist: storage.ghost.io</div><h5><a href="https://www.platformer.news/meta-muse-spark-ai-race/">Meta has a new model</a></h5><div> from <span>@CaseyNewton</span></div></header><div>Nine months after an expensive overhaul, the company says it's back in the AI race — but the race keeps getting faster</div><footer></footer></article></li></ol></footer></blockquote>

Corresponding Notice

Embed this notice
Casey Newton (casey@theforkiverse.com)'s status on Tuesday, 14-Apr-2026 11:02:29 JST Casey Newton
Meta may be back in the AI race — but the race is speeding up, and its rivals' advantages may be compounding. I wrote about Muse Spark: https://www.platformer.news/meta-muse-spark-ai-race/
In conversationabout 2 months ago from theforkiverse.compermalink
Attachments
1. But the frontier has not been waiting for Meta to catch up with it. The day before Meta’s launch, Anthropic announced Claude Mythos Preview, a model so capable at discovering software vulnerabilities that it would not release it to the general public. The flip side of that is that the model is extremely capable at agentic coding — it scored 93.9% on SWE-bench Verified, compared to 77.4% for Muse Spark — and Anthropic is already presumably using it heavily as it builds its next set of models and products. OpenAI will reportedly release a similar model soon, and will also restrict its use. (That model is distinct from another next-generation model, codenamed “Spud,” said to be arriving soon.) Overnight, the true frontier in AI development became a model that’s too good to ship. And because those labs can use those models in development, the gap between them and everyone else may be compounding.
  https://cdn.masto.host/theforkiversecom/media_attachments/files/116/377/632/460/311/715/original/8eed0056927628db.png
2. Domain not in remote thumbnail source whitelist: storage.ghost.io
  Meta has a new model
  from @CaseyNewton
  Nine months after an expensive overhaul, the company says it's back in the AI race — but the race keeps getting faster

Public

Embed Notice

HTML Code

Corresponding Notice