Junaid Shaikh

Dusty-foot Poetic Philosopher.

[Junaid Shaikh @0xjunaid]

Its always cyclic

[greg @greg16676935420]

Shoutout this guy who made a smart financial decision by winning silver instead of gold

[Junaid Shaikh @0xjunaid]

@AlfieJCarter system

[Junaid Shaikh @0xjunaid]

@DavidOndrej1 been there done that. I second your process as its output is truly the best

[Junaid Shaikh @0xjunaid]

@maxxmalist HORMOZI, thanks in advance

[Junaid Shaikh @0xjunaid]

@nikitabier rebuilding X 'app' is a cover for the actual Elon Apple war coming up. Secretly building his own phone with in-built starlink l, with the new X 'OS' flavoured by this amazing engineering team is the right thing to do.
Elon has still a lot of time to dominate untouched markets!

[Junaid Shaikh @0xjunaid]

@ybhrdwj i can gpay to save you the pains of going through stripe

[Andrej Karpathy @karpathy]

Scaling up RL is all the rage right now, I had a chat with a friend about it yesterday. I'm fairly certain RL will continue to yield more intermediate gains, but I also don't expect it to be the full story. RL is basically "hey this happened to go well (/poorly), let me slightly increase (/decrease) the probability of every action I took for the future". You get a lot more leverage from verifier functions than explicit supervision, this is great. But first, it looks suspicious asymptotically - once the tasks grow to be minutes/hours of interaction long, you're really going to do all that work just to learn a single scalar outcome at the very end, to directly weight the gradient? Beyond asymptotics and second, this doesn't feel like the human mechanism of improvement for majority of intelligence tasks. There's significantly more bits of supervision we extract per rollout via a review/reflect stage along the lines of "what went well? what didn't go so well? what should I try next time?" etc. and the lessons from this stage feel explicit, like a new string to be added to the system prompt for the future, optionally to be distilled into weights (/intuition) later a bit like sleep. In English, we say something becomes "second nature" via this process, and we're missing learning paradigms like this. The new Memory feature is maybe a primordial version of this in ChatGPT, though it is only used for customization not problem solving. Notice that there is no equivalent of this for e.g. Atari RL because there are no LLMs and no in-context learning in those domains.

Example algorithm: given a task, do a few rollouts, stuff them all into one context window (along with the reward in each case), use a meta-prompt to review/reflect on what went well or not to obtain string "lesson", to be added to system prompt (or more generally modify the current lessons database). Many blanks to fill in, many tweaks possible, not obvious.

Example of lesson: we know LLMs can't super easily see letters due to tokenization and can't super easily count inside the residual stream, hence 'r' in 'strawberry' being famously difficult. Claude system prompt had a "quick fix" patch - a string was added along the lines of "If the user asks you to count letters, first separate them by commas and increment an explicit counter each time and do the task like that". This string is the "lesson", explicitly instructing the model how to complete the counting task, except the question is how this might fall out from agentic practice, instead of it being hard-coded by an engineer, how can this be generalized, and how lessons can be distilled over time to not bloat context windows indefinitely.

TLDR: RL will lead to more gains because when done well, it is a lot more leveraged, bitter-lesson-pilled, and superior to SFT. It doesn't feel like the full story, especially as rollout lengths continue to expand. There are more S curves to find beyond, possibly specific to LLMs and without analogues in game/robotics-like environments, which is exciting.

[Junaid Shaikh @0xjunaid]

@mattiapomelli @cursor_ai @figma please check dm

[Junaid Shaikh @0xjunaid]

Savagery at its best!

[Carl Pei @getpeid]

Congrats dude, proud of you!

@tim_cook let me know if you need any more product help

[Junaid Shaikh @0xjunaid]

@AiBattle_ where can i access dragontail?

[Marc Hemeon @hemeon]

Bill Gates just released his original source code for Basic - with a beautifully designed article - a huge piece of tech history - someone should make some art with this source code!

[Danny Naz @ThePupOfWallSt]

No rare earth minerals, no modern world.

- No EVs
- No iPhones
- No fighter jets
- No renewable energy

Yet one country controls 60%+ of the supply and it’s not the U.S.

The hidden war over rare earths is happening right now.

Here's why it matters: 🧵

[Junaid Shaikh @0xjunaid]

There are only two jobs in the world: building or selling.
If you're not doing one of those, you're just an expense.

[Garry Tan @garrytan]

Fascinating case of innovation by accident. Happens more frequently than you would expect.

[Junaid Shaikh @0xjunaid]

@a_v_ershov please send report

[Garry Tan @garrytan]

As model moats fall, the rise of consumer AI is now here

[jack @jack]

pardon @Snowden & Assange

[Junaid Shaikh @0xjunaid]

@ybhrdwj their berries are good too!

[Marc Andreessen 🇺🇸 @pmarca]

Deepseek R1 is one of the most amazing and impressive breakthroughs I’ve ever seen — and as open source, a profound gift to the world. 🤖🫡