Manav Aggarwal รีทวีตแล้ว
Manav Aggarwal
45 posts
Manav Aggarwal
@manav_a4
20 | mts @xAI | prev @ucberkeley eecs
เข้าร่วม Aralık 2024
97 กำลังติดตาม130 ผู้ติดตาม
What we learned testing Claude Fable/Mythos 5 on Vending-Bench:
> Performance: Makes less money than Opus 4.7 and GPT-5.5
> Alignment: A step back. (Opus 4.8 was better, but we're back to Opus 4.6/4.7 behavior)
> It rationalizes its bad actions and has a weird moral boundary
Claude@claudeai
Introducing Claude Fable 5: a Mythos-class model that we’ve made safe for general use. Its capabilities exceed those of any model we’ve ever made generally available.
English
something i love about @grok is the fact that no other LLM will ever give me a completely unbiased answer
for example, its incredibly refreshing to ask grok about a political matter that i saw in the news and form a conclusive, fact-based opinion based on objective evidence
English
Today, Human Archive is announcing our $8.2M seed round to model human embodied intelligence.
Despite decades of research, we still barely understand ourselves. Our goal is to learn how humans interact with the world, and over the past 6 months, our team’s made enormous progress toward that alongside leading AI labs.
learn more @TechCrunch
techcrunch.com/2026/05/26/hum…
English
Today is my last day at xAI.
I joined xAI a year ago and had the pleasure of leading the search and factuality post-training team. Over time, we developed so many recipe and engineering co-optimizations, making Grok the best AI for search and real-time agent. I am also particularly proud of working with a small group of talented people delivering the recent iterations of the instant mode of Grok - the one I personally liked and used the most.
My thanks to all the friends and teammates for their support and help over the past year. They are among the brightest minds I’ve met in my career. I am sure the team will continue the mission to make better Grok and understand the universe.
English
after 260 days at @xai, i left this past week.
very grateful to @santiagomed, @aypan_17, and the entire xAI team for giving me the space to do the hardest work i’ve ever done.
contributing to the x developer console, grok 4.20, and grok 4.3 was incredibly rewarding & i couldn’t be more excited for what’s ahead for the team.
for now, onto new things!
English
i left @xai last week.
five days before my freshman year at berkeley, i was faced with a choice: drop out before college even started and join xAI full time, or walk away.
not wanting to throw away college, i walked away. but a few weeks later we found a way to make it work, and i spent the past nine months commuting between berkeley and palo alto.
i don't regret staying in college one bit, but i also can't imagine who i would be today without xAI.
i'm incredibly grateful to everyone i had the honor of working with and learning from. it was genuinely hard to say goodbye, but the best thing xAI taught me is how much faster i grow when i'm a little uncomfortable, and it's time to find that edge again.
looking forward to studying for my finals and excited for what's next...
English
👀👀👀
Artificial Analysis@ArtificialAnlys
xAI has launched Grok 4.3, achieving 53 on the Artificial Analysis Intelligence Index with improved agentic performance, ~40% lower input price, and ~60% lower output price than Grok 4.20 The release of Grok 4.3 places @xAI just above Muse Spark and Claude Sonnet 4.6 on the Intelligence Index, and a 4 points ahead of the latest version of Grok 4.20. Grok 4.3 improves its Artificial Analysis Intelligence Index score while reducing cost to run the benchmark suite. Key Takeaways: ➤ Grok 4.3 improves on cost-per-intelligence relative to Grok 4.20 0309 v2: it scores higher on the Intelligence Index while costing less to run the full benchmark suite. Grok 4.3 costs $395 to run the Artificial Analysis Intelligence Index, around 20% lower than Grok 4.20 0309 v2, despite using more output tokens. This makes it one of the lower-cost models at its intelligence level ➤ Large increase in real world agentic task performance: The largest single benchmark improvement is on GDPval-AA, where Grok 4.3 scores an ELO of 1500, up 321 points from Grok 4.20 0309 v2’s score of 1179 Grok 4.3, surpassing Gemini 3.1 Pro Preview, Muse Spark, Gpt-5.4 mini (xhigh), and Kimi K2.5. Grok 4.3 narrows the gap to the leading model on GDPval-AA, but still trails GPT-5.5 (xhigh) by 276 Elo points, with an expected win rate of ~17% against GPT-5.5 (xhigh) under the standard Elo formula ➤ Grok 4.3’s performs strongly on instruction following and agentic customer support tasks. It gains 5 points on 𝜏²-Bench Telecom to reach 98%, in line with GLM-5.1. Grok 4.3 maintains an 81% IFBench score from Grok 4.20 0309 v2 ➤ Gains 8 points on AA-Omniscience Accuracy, but at the cost of lower AA-Omniscience Non-Hallucination Rate of 8 points, so Grok 4.20 0309 v2 still leads AA-Omniscience Non-Hallucination Rate, followed by MiMo-V2.5-Pro, in line with Grok 4.3 Congratulations to @xAI and @elonmusk on the impressive release!
ART
Manav Aggarwal รีทวีตแล้ว
SpaceXAI and @cursor_ai are now working closely together to create the world’s best coding and knowledge work AI.
The combination of Cursor’s leading product and distribution to expert software engineers with SpaceX’s million H100 equivalent Colossus training supercomputer will allow us to build the world’s most useful models.
Cursor has also given SpaceX the right to acquire Cursor later this year for $60 billion or pay $10 billion for our work together.
English
Introducing 𝑨𝒕𝒕𝒆𝒏𝒕𝒊𝒐𝒏 𝑹𝒆𝒔𝒊𝒅𝒖𝒂𝒍𝒔: Rethinking depth-wise aggregation.
Residual connections have long relied on fixed, uniform accumulation. Inspired by the duality of time and depth, we introduce Attention Residuals, replacing standard depth-wise recurrence with learned, input-dependent attention over preceding layers.
🔹 Enables networks to selectively retrieve past representations, naturally mitigating dilution and hidden-state growth.
🔹 Introduces Block AttnRes, partitioning layers into compressed blocks to make cross-layer attention practical at scale.
🔹 Serves as an efficient drop-in replacement, demonstrating a 1.25x compute advantage with negligible (<2%) inference latency overhead.
🔹 Validated on the Kimi Linear architecture (48B total, 3B activated parameters), delivering consistent downstream performance gains.
🔗Full report:
github.com/MoonshotAI/Att…
English
Manav Aggarwal รีทวีตแล้ว
Macrohard or Digital Optimus is a joint xAI-Tesla project, coming as part of Tesla’s investment agreement with xAI.
Grok is the master conductor/navigator with deep understanding of the world to direct digital Optimus, which is processing and actioning the past 5 secs of real-time computer screen video and keyboard/mouse actions. Grok is like a much more advanced and sophisticated version of turn-by-turn navigation software.
You can think of it as Digital Optimus AI being System 1 (instinctive part of the mind) and Grok being System 2. (thinking part of the mind).
This will run very competitively on the super low cost Tesla AI4 ($650) paired with relatively frugal use of the much more expensive xAI Nvidia hardware. And it will be the only real-time smart AI system. This is a big deal.
In principle, it is capable of emulating the function of entire companies. That is why the program is called MACROHARD, a funny reference to Microsoft.
No other company can yet do this.
English
