HackerTrans
TopNewTrendsCommentsPastAskShowJobs

MukundMohanK

no profile record

Submissions

Barbie's new pink (insulin) pumps – help children with type 1 diabetes

science.org
2 points·by MukundMohanK·12 mesi fa·0 comments

A new Mongolian tyrannosauroid and the evolution of Eutyrannosauria

nature.com
2 points·by MukundMohanK·anno scorso·0 comments

Physicists on a remote island: we visit the ultimate quantum party

nature.com
2 points·by MukundMohanK·anno scorso·0 comments

NASA reveals monster star clumps in galactic wreckage

sciencedaily.com
3 points·by MukundMohanK·anno scorso·1 comments

LLMs Can Write Efficient CUDA Kernels

arxiv.org
2 points·by MukundMohanK·anno scorso·0 comments

Mixed-Chip Clusters Enable Efficient Large-Scale AI Training

arxiv.org
3 points·by MukundMohanK·anno scorso·0 comments

I don't think AGI is right around the corner

dwarkesh.com
2 points·by MukundMohanK·anno scorso·0 comments

How old are the Dead Sea Scrolls? An AI model can help

economist.com
2 points·by MukundMohanK·anno scorso·0 comments

Context is all you need

fakepixels.substack.com
2 points·by MukundMohanK·anno scorso·0 comments

Silicon Valley wants to help me make a superbaby. Should I let it?

sfstandard.com
1 points·by MukundMohanK·anno scorso·1 comments

[untitled]

1 points·by MukundMohanK·anno scorso·0 comments

[untitled]

1 points·by MukundMohanK·anno scorso·0 comments

Ivy Derangement Syndrome – Immolation from the Eyes of a Harvard Professor

nytimes.com
2 points·by MukundMohanK·anno scorso·0 comments

Academic project turns into a $600M crowdsourced LLM benchmarking startup

bloomberg.com
2 points·by MukundMohanK·anno scorso·0 comments

comments

MukundMohanK
·12 mesi fa·discuss
Re:domain-specific solutions inevitably outcompete general-purpose ones. Certainly not the case for coding agents.

Vertical software won because GUIs were scarce. General LLMs are winning because context is abundant.
MukundMohanK
·anno scorso·discuss
Reality is here whether we like it or not - https://fred.stlouisfed.org/graph/?g=1DEP0
MukundMohanK
·anno scorso·discuss
Between last April and now, swe-bench scores have gone up from 25%-70%.

Sure, they're being overfitted to the dataset. But with most performing similarly across even the hardest of 3rd party benchmarks, think frontier math back in Nov and now, we're closer than ever to a specialisation shift.

Hard to say at what % but once code reviews get better its likely 2025 is the last year SWE is a sought after job * demand and supply both