S. M. Hossein Hamidi

Swimming upstream, thinking critically and learning relentlessly.

[S. M. Hossein Hamidi @smhhamidi]

وَ اجْعَلْ قَلْبِی وَاثِقاً بِمَا عِنْدَكَ، وَ هَمِّی مُسْتَفْرَغاً لِمَا هُوَ لَكَ، وَ اسْتَعْمِلْنِی بِمَا تَسْتَعْمِلُ بِهِ خَالِصَتَكَ
https://t.co/Xhkw5heZEu

[Arnaud Bertrand @RnaudBertrand]

I'm probably going to get a lot of hate for this but Altman is right here, that's the right way of seeing AI models: AI is a general purpose technology and models will end up being a utility like electricity.

In fact I'm somewhat surprised he's admitting to this because it makes companies like OpenAI or Anthropic a lot less valuable: it means they'll become mere commodities, much like telecom companies or electricity providers are.

The real value will lie in the application layer - what you actually DO with AI - as opposed to the models themselves. Much like the real valuable companies enabled by the internet weren't the telecom companies but businesses like Google, Amazon or Alibaba.

I actually wrote a whole article explaining exactly this last month titled "There is no AI race": https://t.co/48ifTX1YsD

[S. M. Hossein Hamidi @smhhamidi]

"technology is economy, and economy is technology"

[signüll @signulll]

apple products are sold out. infinite demand for inference, rate limits everywhere. anthropic & openai growing at ridiculous pace. nvidia & google at all time highs. record meta earnings.

we are now more supply constrained than demand constrained. i suspect we’ll see inflationary pressure again very soon.

fundamentally, technology is the economy, & the economy is technology.

[Paul Graham @paulg]

You can't get the same effect from the "Innovation Center" set up by the provincial government. You have to fill out forms to get space there, it's locked at night, and you can't make holes in the walls. You might even have to deal with bureaucrats hired to "help" you.

[S. M. Hossein Hamidi @smhhamidi]

@rezabakhtyari5 ناظر به مصداق خاصی نبود و نیست.

[S. M. Hossein Hamidi @smhhamidi]

گفت:
تا بند ناف پاره نشود، شیر تو سینه‌های مادر نمی‌آید.

[S. M. Hossein Hamidi @smhhamidi]

@rezabakhtyari5 دستاوردها، محصول انقطاع و دل بریدن‌ها است.

[Arnaud Bertrand @RnaudBertrand]

I have a website about Traditional Chinese Medicine that I spent literal years building. When I asked questions to Claude about the topic, it parroted almost word-for-word what I myself wrote.

So please spare us the gaslighting about training AI on others' work...

[S. M. Hossein Hamidi @smhhamidi]

چون وا نمی کنی گرهی، خود گره مشو
ابرو گشاده باش چو دستت گشاده نیست
https://t.co/O9VcAHbDCD

[S. M. Hossein Hamidi @smhhamidi]

اللّهُمَّ إِنِّي أَسْأَلُكَ أَنْ تَجْعَلَ لِي إِلى كُلِّ خَيْرٍ سَبِيلاً،
وَمِنْ كُلِّ ما لا تُحِبُّ مانِعاً

https://t.co/aIAVqozNNI

[S. M. Hossein Hamidi @smhhamidi]

@hajikarimi @mahdikiani متوجه نشدم.
چرا پرسیدن، مگر وارد حیطه‌ی استدلال شدن نیست؟

[S. M. Hossein Hamidi @smhhamidi]

@goddek https://t.co/LW6lsRR5Kv

[Chinese Embassy in US @ChineseEmbinUS]

🔴 #China is deeply shocked by and strongly condemns the #US’s blatant use of force against a sovereign state and action against its president. Such hegemonic acts of the U.S. seriously violate international law and #Venezuela 's sovereignty, and threaten peace and security in Latin America and the Caribbean region. China firmly opposes it. We call on the U.S. to abide by international law and the purposes and principles of the UN Charter, and stop violating other countries’ sovereignty and security.

[Jaana Dogan ヤナ ドガン @rakyll]

Building at large companies is hard right now due to the legacy infrastructure. 10-15 years ago, some of these companies were building in house to be frontier. Today, what's available internally is too disconnected from what's available outside and is often years behind.

[Paul Graham @paulg]

The stupidity of this regulation will be revealed gradually as technology evolves.

[andrew chen @andrewchen]

1/ Many of the biggest implosions in recent history - especially ecommerce - have been due to startups getting addicted to paid marketing while fooling themselves on Customer Acqusition Costs. As spend scales, it always gets more expensive and harder to track - never less.

[Aakash Gupta @aakashgupta]

This obscures what actually happened here.

In 2013, Elon warned Demis Hassabis that the future of AI shouldn’t be controlled by Larry Page. He tried to cobble together financing with PayPal co-founder Luke Nosek to buy DeepMind himself. Google closed for $500M anyway.

Losing that deal is why OpenAI exists. Musk co-founded it in 2015 specifically to break Google’s grip on AI talent. Then he poached Ilya Sutskever from Google Brain with $1.9M salary plus signing bonus to become OpenAI’s chief scientist. Page “refused to hang out” with Musk after that.

So yes, five people shaped 30 years of tech. But the more interesting pattern: those five people spent most of that time trying to prevent each other from winning. Zuck’s $800M bid for DeepMind failed. Elon’s counter-bid failed. Both responded by building competing AI labs.

The jet story sounds like networking. What it actually shows is surveillance. Every major tech founder in 2013 was tracking who controlled the best AI researchers and trying to either acquire them or block rivals from acquiring them.

The same handful of people keep appearing because there are only about 200 researchers in the world capable of building frontier AI systems, and every billionaire in tech knew it a decade before the rest of us figured it out.

[Aakash Gupta @aakashgupta]

The entire robotics industry is about to compress a decade of progress into 18 months, and nobody’s pricing it in.

The hardware has been ready for years. Boston Dynamics had Atlas doing backflips in 2018. The bottleneck was never motors or actuators. It was that every robot behavior had to be hand-coded. Pick up a box? That’s one program. Pick up a bottle? Different program. Move the box from shelf A to shelf B in a warehouse with slightly different lighting? Start over.

Foundation models broke this completely.

Before VLAs, teaching a robot one skill gave you exactly one skill. Zero compounding. Zero transfer. A robot trained to fold shirts couldn’t fold towels without starting from scratch. The labor intensity of data generation meant robotics datasets stayed narrow, robots overfit, and small variations like object weight or table height caused failures.

Now a single Gemini Robotics model handles tasks it has never seen in training. Google’s On-Device model learns new behaviors with 50-100 demonstrations. Not 50,000. Fifty. That’s a 1000x reduction in the data requirement for new capabilities.

The speed implications cascade through everything.

First order: deployment timelines collapse. What took robotics teams 6-12 months of custom programming now takes days of fine-tuning. Second order: the addressable market explodes. Tasks that were never economical to automate suddenly are, because the integration cost dropped by orders of magnitude. Third order: the data flywheel accelerates. Every robot running Gemini Robotics feeds learning back into the foundation model. More deployments means faster improvement means more deployments.

Physical Intelligence raised at $2.4B because investors finally understood this. Boston Dynamics partnered with Toyota Research Institute to bolt Large Behavior Models onto Atlas. Every humanoid company is scrambling to either build or license the intelligence layer they don’t have.

The market is still valuing robotics companies on their hardware differentiation. But hardware is commoditizing. Boston Dynamics spent a decade perfecting locomotion, and now that’s table stakes. The value is migrating entirely to whoever owns the foundation model that generalizes across embodiments.

Google trained Gemini on the largest multimodal corpus ever assembled. Then they added physical actions as an output modality. That’s not a robotics company bolting on AI. That’s an AI company whose models now output motor commands.

The companies pricing this correctly are building around foundation model access, not around proprietary hardware. The companies pricing this wrong are still acting like the moat is in the mechanical engineering.

AGI moving into the physical world isn’t a 10-year prediction. Gemini Robotics shipped in March. The 1.5 version with chain-of-thought reasoning shipped in September. They’re iterating on a 6-month release cycle while hardware companies iterate on 3-year cycles.

The gap between software intelligence timelines and hardware development timelines is the entire trade.

[Sumit Kumar @TweetsOfSumit]

Still daily business in German enterprises.

If you have an “agile coach” anywhere near your place in the org chart, leave the company. Thank me later.

[Paul Graham @paulg]

A startup told me a highly-funded competitor bought the .com of their name. I told them this is good news in a way. It means the competitor is (a) amateurish and (b) afraid of them, both of which suggest the competitor will lose in the long term.

[Paul Graham @paulg]

To bad designers, great designs start to look outdated after a couple decades simply because they're associated with a time in the past. If you're not careful, they'll "update" them.