Nanocode: The best Claude Code that $200 can buy in pure JAX on TPUs
by desideratum on 4/5/2026, 2:21:17 PM
https://github.com/salmanmohammadi/nanocode/discussions/1
Comments
by: wwfn
Tangential (but topical in that "The threat is comfortable drift toward not understanding what you're doing" is also on the front page):<p>Is the generated python code in the example wrong?<p>The prompt<p>> Develop a Python function that removes any falsey values from a list. Return the modified list without creating a new one.<p>Is answered with list comprehension, which makes a new list and leaves the original unmodified (never mind that the *args input necessarily can't be a modifiable list?)<p><pre><code> def remove_falsey_values(*args): return [val for val in args if val] </code></pre> Whereas I'd expect something like<p><pre><code> def remove_falsey_values(l): for i in reversed(range(len(l))): if not l[i]: l.pop(i) # returned list is linked to input l return l a = [1, 0, False, 'foo'] x = remove_falsey_values(a) x[0] = 2 print(a) # [2,'foo']</code></pre>
4/5/2026, 5:11:02 PM
by: jaboostin
As someone with zero ML experience, this was a super interesting and digestible read!
4/5/2026, 4:40:27 PM
by: bdbdbdb
Dumb question - and I'm not trying diminish the achievement here, I just genuinely don't understand:<p>Why would people want to spend $200 to train a coding model when there are free coding models?
4/5/2026, 4:21:05 PM