Ibrahim Sharaf

ml/nlp @Cision ms @UW

[Ibrahim Sharaf @IbrahimSharraf]

@amasad Not anymore, you cant 😭😂

[Richard Sutton @RichardSSutton]

The bitter lesson in 26 words:

Don’t be distracted by human knowledge, as AI has been historically.
Instead focus on methods for creating knowledge that scale with computation, like search and learning.

[kitze the 🐐 @thekitze]

> be github
> invent copilot
> you are literally the first one
> you are literally the only one
> you literally have access to all the code in the world
> get mogged by literally every single agentic bs that came out in the past few years

this level of fumble should be studied

[Sara Hooker @sarahookr]

A really excellent book.

A few people independently told me this was one of their favorite books over a decade ago. I bought it, and it became one of the textbooks on my shelf I revisit from time to time to spark the joy of holding ideas to a different light.

It brings to life the elegance of information theory.

A good day to recognize 10 years from the passing of David McKay.

[MOHAMMED𓂆 @Mo7mmd30]

Everyone felt sad for the penguin walking alone and the monkey rejected by his mother. But this video is far more heartbreaking, yet it didn’t receive the same attention.

[Paul Graham @paulg]

Prediction: In the AI age, taste will become even more important. When anyone can make anything, the big differentiator is what you choose to make.

https://t.co/3GQUlfH58t

[Ibrahim Sharaf @IbrahimSharraf]

@AtnsXBT Love this, i did a similar thing in siracusa, sicily. Such a profound trip.

[Andrej Karpathy @karpathy]

I tried to use it this way and basically failed, the models aren't at the level where they can productively iterate on nanochat in an open-ended way. (Though one of the primary motivations for me writing nanochat is that I'd very much love for it to be used this way as a benchmark for agents, and I'd love it if it worked over time). I'm open to this just being skill issue.

E.g. here some of the things I'd be suspicious about:
- the zoo of torch compile flags can knowingly be abused to get +1% gains but often at the cost of +30min compile time. This is why modded-nanogpt prohibits torch compile kwarg engineering and why I haven't done any in nanochat either. i wouldn't reliably expect the model to notice, consider, or flag this kind of an issue or seek clarification.
- ns_steps=3 might be a tiny bit of speed, but does the model also volunteer to make sure quality doesn't fall too much?
- same thing for deleting .float() cast - sure you can remove it and get VRAM/speed gains but it's there for a clear reason (extra precision in the loss function). Removing it means you absolutely have to make sure that the lower precision is ok validation loss wise, in a highly controlled experiment.

Overall I'm still struggling with getting the models to do significantly more basic things. For example, Opus keeps incorrectly "cleaning up" my comments when it doesn't understand them even when it's completely unrelated to the task, rude! It keeps violating and ignoring CLAUDE .md instructions on coding style but when I ask, it correctly points out all the violations. I know, I'm supposed to be using some kind of a /cleanup. Yesterday it gave me a table of results and incorrectly reported which experiment worked best (the table showed xyz=20 was best and it incorrectly claimed that xyz=12 was). Basically - much simpler things still fail routinely than something open-ended like "improve nanochat". (I've been doing a lot of YELLING IN UPPER CASE and I think this could actually be a really good metric for A/B testing instead of the inline survey thing.). Still incredibly net useful with oversight and with clear, well-scoped tasks.

I definitely haven't given up on automatic closed-loop experiments with the models. It would be so glorious. I had 2 iterations that basically didn't work but I have ideas for the 3rd.

[Aakash Gupta @aakashgupta]

Sounds incredible until you read the fine print. The compiler generates less efficient code than GCC with all optimizations disabled. It doesn’t have its own assembler or linker. It can’t produce a 16-bit x86 code generator. And Carlini himself says it has “nearly reached the limits of Opus’s abilities.” New features and bugfixes kept breaking existing functionality.

So what did $20,000 and two weeks actually buy? A compiler that passes 99% of GCC’s torture tests but can’t match the output quality of a tool that’s had 37 years of human engineering. That’s the constraint nobody’s pricing in.

The real story is in the cost curve, not the capability demo. $20,000 for 100,000 lines means $0.20 per line of generated code. A senior compiler engineer costs roughly $150/hour. At maybe 50 polished lines per hour for something this complex, that’s $3/line. AI just did it at 15x cheaper, and it will only get cheaper from here.

But the code isn’t equivalent. The AI version needs a human to finish the assembler, fix the linker, optimize the output, and prevent regressions. Those are the hardest 20% of the problem, and they represent 80% of the engineering value. Anthropic built the demo. Shipping the product still requires humans.

This tells you exactly where we are in the autonomous software timeline. AI can now produce impressive first drafts of complex systems at trivial cost. Turning those drafts into production software still requires the judgment that costs $300K+ per year in compiler engineer salary. The gap between “compiles the Linux kernel” and “replaces GCC” is measured in decades of accumulated engineering wisdom that no model has internalized yet.

The companies that understand this will use agent teams to generate the 80% and hire engineers to finish the 20%. The companies that don’t will ship $20,000 compilers that produce slower code than a free tool from 1987.

[Ibrahim Sharaf @IbrahimSharraf]

@DearS_o_n @grok why

[Valerie Zink @valeriezink]

I can’t in good conscience continue to work for Reuters given their betrayal of journalists in Gaza and culpability in the assassination of 245 our colleagues.

[Ibrahim Sharaf @IbrahimSharraf]

@drsamehmesallum @grok وضع تيران و صنافير وصل لايه دلوقتي

[Barry Malone @malonebarry]

Mariam Dagga, killed by Israel today, was an incredibly talented photographer. She relentlessly and courageously documented the genocide. I'm going to share some of her images. Graphic content included.

[Zarah Sultana MP @zarahsultana]

Legacy media are not our allies. They are the mouthpiece of the ruling class.

And how many of them have spoken out against Israel’s genocide in Gaza and the direct targeting of Palestinian journalists?

You cannot give these people an inch.

Their smears won’t work this time.

[Andrej Karpathy @karpathy]

I am (slowly) re-reading the Tolkien legendarium (of which Lord of the Rings is a small part). The whole body of work is so incredible and there's nothing else like it... it dilutes other worlds of fiction. Wait - your story doesn't have a comprehensive history/mythology spanning multiple ages all the way back to a creation myth as detailed in separate volumes? You didn't first invent new languages and dialects for your characters? You didn't pack it with powerful themes and stories written it in a beautiful, archaic style and compose poems and songs alongside? It didn't take you multiple decades of iteration? And what of all the uncharted territory still remaining? Is Tom Bombadil one of the Ainur. Where are the Entwives. What happened to the two unaccounted Istari. Can we hear more about what it was like in Cuiviénen when the elves first awoke? Or to see the light of the two trees of Valinor. Or of the splendor of the caves of Aglarond.

What's most on my mind though - the Tolkien legendarium is imo a concrete example of a height of culture. Does AI, today or soon, make it easier to reach this high via empowerment in both writing and ideation? Or harder, when quick wins are tempting and ~free, and an independent ability to create is stifled. If such a body of work is made again but now with heavy AI assistance, does it inspire the same wonder? What if thousands of them come out on demand with just a prompt? Why do you feel cheated when you learn that something your read was AI generated? Is it transient or a function of capability? Is it slop? What is slop? Or is wonder inseparable from its own creation myth of a lifelong obsession of a mind like your own? So many questions.

[أنس الشريف Anas Al-Sharif @AnasAlSharif0]

هذه وصيّتي، ورسالتي الأخيرة.
إن وصلَتكم كلماتي هذه، فاعلموا أن إسرائيل قد نجحت في قتلي وإسكات صوتي.
بداية السلام عليكم ورحمة الله وبركاته

يعلم الله أنني بذلت كل ما أملك من جهدٍ وقوة، لأكون سندًا وصوتًا لأبناء شعبي، مذ فتحت عيني على الحياة في أزقّة وحارات مخيّم جباليا للاجئين، وكان أملي أن يمدّ الله في عمري حتى أعود مع أهلي وأحبّتي إلى بلدتنا الأصلية عسقلان المحتلة "المجدل" لكن مشيئة الله كانت أسبق، وحكمه نافذ.
عشتُ الألم بكل تفاصيله، وذُقت الوجع والفقد مرارًا، ورغم ذلك لم أتوانَ يومًا عن نقل الحقيقة كما هي، بلا تزوير أو تحريف، عسى أن يكون الله شاهدًا على من سكتوا ومن قبلوا بقتلنا، ومن حاصروا أنفاسنا ولم تُحرّك أشلاء أطفالنا ونسائنا في قلوبهم ساكنًا ولم يُوقِفوا المذبحة التي يتعرّض لها شعبنا منذ أكثر من عام ونصف.

أوصيكم بفلسطين، درةَ تاجِ المسلمين، ونبضَ قلبِ كلِّ حرٍّ في هذا العالم.
أوصيكم بأهلها، وبأطفالها المظلومين الصغار، الذين لم يُمهلهم العُمرُ ليحلموا ويعيشوا في أمانٍ وسلام،
فقد سُحِقَت أجسادهم الطاهرة بآلاف الأطنان من القنابل والصواريخ الإسرائيلية، فتمزّقت، وتبعثرت أشلاؤهم على الجدران.

أوصيكم ألّا تُسكتكم القيود، ولا تُقعِدكم الحدود، وكونوا جسورًا نحو تحرير البلاد والعباد، حتى تشرق شمسُ الكرامة والحرية على بلادنا السليبة.
أُوصيكم بأهلي خيرًا،
أوصيكم بقُرّة عيني، ابنتي الحبيبة شام، التي لم تسعفني الأيّام لأراها تكبر كما كنتُ أحلم.

وأوصيكم بابني الغالي صلاح، الذي تمنيت أن أكون له عونًا ورفيق دربٍ حتى يشتدّ عوده، فيحمل عني الهمّ، ويُكمل الرسالة.

أوصيكم بوالدتي الحبيبة، التي ببركة دعائها وصلتُ لما وصلت إليه، وكانت دعواتها حصني، ونورها طريقي.
أدعو الله أن يُربط على قلبها، ويجزيها عنّي خير الجزاء.

وأوصيكم كذلك برفيقة العمر، زوجتي الحبيبة أم صلاح بيان، التي فرّقتنا الحرب لأيامٍ وشهورٍ طويلة، لكنها بقيت على العهد، ثابتة كجذع زيتونة لا ينحني، صابرة محتسبة، حملت الأمانة في غيابي بكلّ قوّة وإيمان.

أوصيكم أن تلتفوا حولهم، وأن تكونوا لهم سندًا بعد الله عز وجل.
إن متُّ، فإنني أموت ثابتًا على المبدأ، وأُشهد الله أني راضٍ بقضائه، مؤمنٌ بلقائه، ومتيقّن أن ما عند الله خيرٌ وأبقى.

اللهم تقبّلني في الشهداء، واغفر لي ما تقدّم من ذنبي وما تأخّر، واجعل دمي نورًا يُضيء درب الحرية لشعبي وأهلي.
سامحوني إن قصّرت، وادعوا لي بالرحمة، فإني مضيتُ على العهد، ولم أُغيّر ولم أُبدّل.

لا تنسوا غزة…
ولا تنسوني من صالح دعائكم بالمغفرة والقبول.
أنس جمال الشريف
06.04.2025

هذا ما أوصى بنشره الحبيب الغالي أنس عند استشهاده.
إدارة الصفحة

[Nathan C. Frey @nc_frey]

The AI4Science workshop @NeurIPSConf is hosting a dataset proposal competition with a $10K USD prize pool!

Submit a 2 page proposal for a dataset you'd like to create and why.

Thanks to @AdaFang_ & the organizers for making this happen!

More info and link below👇

[Andre @Youkhna]

Built this from scratch.
No team of 10,000. No venture funding.
Just a stubborn idea and a lot of late nights.

Physical keyboard. E-Paper Display.

[Deep Learning Indaba @DeepIndaba]

📣 The official press release for #DLI2025 is here!
We’re heading to Kigali to connect 1000+ brilliant minds shaping the future of AI in Africa 🌍
Read more about what’s coming:
🔗https://t.co/lvv6Huue41

#Indaba2025
#Urunana