Shaaban

Obligatory opinions are my own disclaimer

[Jason Droege @jdroege]

https://t.co/05e7R1SKmQ

[Alexandr Wang @alexandr_wang]

1/ today we're releasing muse spark, the first model from MSL. nine months ago we rebuilt our ai stack from scratch. new infrastructure, new architecture, new data pipelines. muse spark is the result of that work, and now it powers meta ai. 🧵

[J.J. McCullough @JJ_McCullough]

The wild thing about the "scandal" of the Air Canada CEO is that his crime was purely ideological, as enforced by politicians and media. He broke a "taboo" in a literal sense — something that wasn't of any practical importance, but had spiritual significance to the elite.

[Lucy Hargreaves @lucyhargreaves4]

it doesn’t have to be this way

[Dave West @davidallenwest]

🇨🇦 - CBC are traitors to democracy

[ℏεsam @Hesamation]

everyone has been dunking on her,
“ShE’s tHe HeaD of SafeTy At Meta”
or “wHy did She PoSt tHis?”

let’s talk about this:
> doing research in safety doesn’t mean you’re incapable of making mistakes
> 1000s of OpenClaw users made the exact same mistake as her.
> she was the only victim, it wasn’t a corporate mistake affecting users
> maybe we should talk about the preachers who don’t warn enough about it? if you talk about making billions with OpenClaw maybe you should also share what you did to make sure it’s secure.

the most stupid take on this was “why did she openly share this? it’s bad PR” well if you’re a safety and alignment lead and have AI expert connections on social media, this is exactly what you need to post and raise awareness about, not your dinner pics, even if you take the hit for it.

[Alexandr Wang @alexandr_wang]

New, very needed benchmark from @scale_AI: SWE-Bench Pro

Includes:
- Multi-file edits
- 100+ lines changed on average
- Complex dependencies across large codebases

Current top model scores:
- GPT-5: 23.3%
- Claude Opus 4.1: 22.7%
- Others drop further (<15%)

[Will Kinney @WKCosmo]

PhD cosmologist here. Only about 4% of the universe is.

[Shaaban @MoShaaban]

@perplexity_ai phone app having notifications on by default and sending me a ton of completely unnecessary notifications like it’s some news app is why I went from a lifer to deleting the app 🤷

Yes I know I can just take away notification perms but that’s not the point.

[Shaaban @MoShaaban]

San Francisco is a third world country packed with the highest density of human talent.

[Shaaban @MoShaaban]

@Andercot Yes, off the grid SMRs at northern Canadian lakes with each SMR directly powering a massive compute farm.

[Shaaban @MoShaaban]

Amazing opportunity for anyone interested in AI adoption and AI safety! Feel free to DM me.

[Summer Yue @summeryue0]

🤖 AI agents are crossing into the real world.
But when they act independently—who’s watching?

At Scale, we’re building Agent Oversight: a platform to monitor, intervene, and align autonomous AI.

We’re hiring engineers (SF/NYC) to tackle one of the most urgent problems in AI. Apply here or DM me →
SWE: https://t.co/oJJArT6nEl
FDE: https://t.co/gcxQKxF9bz

[Alexandr Wang @alexandr_wang]

🚨 @OpenAI has launched o3 and o4-mini! 🎉

o3 is absolutely dominating the SEAL leaderboard with #1 rankings in:

🥇: HLE
🥇: Multichallenge (multi-turn)
🥇: MASK (honesty under pressure)
🥇: ENIGMA (puzzle solving)

Congrats @sama @markchen90 & team

🔗: https://t.co/pVIgk6rIcL

[Alexandr Wang @alexandr_wang]

🚨 Gemini 2.5 Pro Exp dropped and it's now #1 across SEAL leaderboards:

🥇 Humanity’s Last Exam
🥇 VISTA (multimodal)
🥇 (tie) Tool Use
🥇 (tie) MultiChallenge (multi-turn)
🥉 (tie) Enigma (puzzles)

Congrats to @demishassabis @sundarpichai & team!

🔗 https://t.co/pVIgk6rIcL

[J.J. McCullough @JJ_McCullough]

Canada will not "diversify" its partners and become some new global power. This has been tried a million times and it never works. Countries grow rich by exploiting their comparative advantage, and Canada's is closeness with the US. Losing that means a poorer Canada, period.

[Shaaban @MoShaaban]

My wife asked me to explain the @luka7doncic and @AntDavis23 trade to her. So here was my best attempt at an analogy

1.Imagine if music worked in multi artist teams.
2.@taylorswift13 is on a team, and she’s got off the the peak of her Eras Tour.
3.Her team trades her to another team for Doja Cat, who is great but also has less time left on her career.
4.Instead of publicly announcing they’re open to trading Taylor to see what the best offer they can get is, her team just directly offers her for Doja.
5.this is a huge miss because other teams would have offered basically anything for Taylor, deals that involve both Kendrick Lamar and SZA, or even Billie Eilish plus a legendary retiring artist like Paul McCartney would have been on the table.
6.The fans of the team that got Taylor are obviously thrilled, while her original team is devastated—they even held a symbolic funeral for their team.
7.Meanwhile, Doja’s original team is celebrating because they got Taylor Swift for such a low price.
https://t.co/sF0Kd7SIvW, Taylor is on the same team as Michael Jackson who is planning to retire in a year or two so she is set to train under him for two years before launching her biggest tour ever in 2027.
9.There’s a conspiracy that the NBA forced this trade to boost viewership, because if Taylor Swift turns into the next Michael Jackson, ratings will skyrocket.
10.The craziest part? Neither Taylor, Doja, nor Michael Jackson asked for this trade—they all found out about the trade after it was done.

[Alexandr Wang @alexandr_wang]

Scale AI & CAIS are releasing Humanity’s Last Exam (HLE), a dataset of 3,000 questions developed with hundreds of subject matter experts (PhDs, profs, etc.) to capture the human frontier of knowledge and reasoning.

State-of-the-art AIs get <10% and are highly overconfident.
🧵

[Shaaban @MoShaaban]

@Andercot This is actually such an insanely excellent analogy !

[Shaaban @MoShaaban]

@hardaway_cam Is this an official angel request 😏

[Shaaban @MoShaaban]

Like every other 30 year old, I’m going to try this blogging thing:

https://t.co/FFdrE3YH1D