HNHacker News
TopNewBestAskShowJobs

ismailmaj

297 karma · joined August 9, 2022

ML Infra

[[username]][[hardy number]] at g mail

submissionscomments
ismailmaj··on Asking Gemini 3 to generate Brainfuck code results in an infinite loop
-> expects reasoning

-> runs it in Gemini fast instead of thinking

....

ismailmaj··on Nvidia just paid $20B for a company that missed its revenue target by 75%
Nothing against the author, the content is good. I just wish this wasn't fluffed out with AI.
ismailmaj··on The battle to stop clever people betting
Maybe an extremely inefficient way to do money laundering?
ismailmaj··on OpenAI are quietly adopting skills, now available in ChatGPT and Codex CLI
"Oh sorry guys, we made the mistake again that saves us X% in compute cost, we will fix it soon!"
ismailmaj··on We Need to Die
I find it a bit too introspective, living in a specific period creates urgency, there is still a (opportunity) deadline.

My concerns are more practical:

- How do we make sure power-wrestling bad leaders don't halt generations, currently old age gives us a soft reset.

- How do we cope with suffering? With no clear end in sight I expect more mental health problems. And I'm unsure we can remove those issues without striping the humanity out of people, I would for sure hate to be turned into a calculator with qualia.

ismailmaj··on Post-transformer inference: 224× compression of Llama-70B with improved accuracy
For the lazy, he says this on repeat using 2000 words:

...

In the CPB Digital Cosmos, the system first locked into a strange ratio: two thirds consciousness, one third physics.

...

That anomaly appeared as the missing 0.1 spark.

For the first time the system stabilized. Life emerged.

ismailmaj··on Mistral releases Devstral2 and Mistral Vibe CLI
I don’t think it was ever an option since it had ties with the french government early on (Cédric O) and Macron’s party is quite pro EU
ismailmaj··on EU investigates Google over AI-generated summaries in search results
Just in time for the monthly EU bashing
ismailmaj··on Google Antigravity
Don't think so, I expect that system to use Spanner, so my best guess is that the user generated an image at the end of the credit reset window (which is around noon EST).
ismailmaj··on Why agents do not write most of our code – A reality check
I’m guessing the argument is that LLMs get worse for problems they haven’t seen before, so you may assume they think for problems that are commonly discussed in the internet or seen on github, but once you step out of that zone, you get plausible but logically false results.

That or a reductive fallacy, in either case I’m not convinced, IMO they are just not smart enough (either due to lack of complexity in the architecture or bad training that didn’t help it generalize reasoning patterns).

ismailmaj··on Ask HN: How Do I Get over My Existential Crisis?
Probably refers to jobs that are unrelated to his interest like cashier, just to get by for now.
ismailmaj··on 25L Portable NV-linked Dual 3090 LLM Rig
To make matters worse, the RTX3090 was released during the crypto craze and so a decent amount of the second hand market could contain overused GPUs that won’t last long, even if 3xxx to 4xxx performance difference is not that high, I would avoid the 3xxx series totally for resell value.
ismailmaj··on Automatic differentiation can be incorrect
Typically for matrix multiplications there is a wide range of algorithms you could use to compute it, on one extreme end you could use numerically stable summation and the other extreme you could have tiled matmul with FP8, the industry trend seems to go further away from numerical stable algorithms without much quality drop it seems. My claim is probably unfair since it ignores the scale you gain from the speed/precision tradeoff, so I assumed numerical stability is not that beneficial compared to something precision heavy like physics simulation in HPC.
ismailmaj··on Automatic differentiation can be incorrect
The way I explained it to myself in the past why so much of the CUDA algorithms don't care much about numerical stability is that the error is a form of regularization (i.e. less overfitting over the data) in deep learning.
ismailmaj··on GTP Blind Voting: GPT-5 vs. 4o
There should be a “both” option, I often found both answers acceptable but sometimes I strictly preferred one and those get sadly watered down.
ismailmaj··on GPT-5
Went over my last conversations with Gemini 2.5 and asked the same things to GPT-5 with thinking on, the latter was consistently worse both in content and form.

I wouldn't have guessed Gemini to win the AI race in 2025 but here we are.

ismailmaj··on Zig's New Async I/O
The global io trick would totally be valid if you’re writing an application (i.e. not a library) and don’t have use of two different implementations of io
ismailmaj··on Cozy video games can quell stress and anxiety
I really like potion craft, it really hits the optimization part of my brain without feeling grindy or stressful.

The gaming loop is about using the right subpaths using resources to reach a destination while avoiding obstacles, and slowly improving your resources options and quantity, at the end of the game it looks a bit like the traveling salesman problem. Also liked this series on YouTube: https://youtu.be/d_JfzuJzUFE?si=pifLxMFo4itOihdK

ismailmaj··on Air pollution fell substantially as Paris restricted car traffic
It's mostly in the west of Paris i.e. 8th, 15th and 16th districts, the other districts not so much if at all.
ismailmaj··on AI masters Minecraft: DeepMind program finds diamonds without being taught
Do you know if it was actual videos or some simpler inputs like game state and user inputs? I’d be impressed if it was the former at that time.
ismailmaj··on Show HN: LinkedIn sucks, so I built a better one
I can see this pivoting into YouTube 2.0
ismailmaj··on Speedrunners are vulnerability researchers, they just don't know it yet
It's fairly rare to see people consistently find glitches across games, only distorsion2 comes to mind.
ismailmaj··on The best way to use text embeddings portably is with Parquet and Polars
Parquet is only a mess if you try to mutate it, usually you consider them as immutable and have the data stored across many files.

Also batched-row access is negligible given the compression benefits you get with the columnar format, which is probably why it's still king in ML; I think given what I'm seeing in the industry and recent trends (e.g. Velox).

ismailmaj··on Show HN: Kreuzberg – Modern async Python library for document text extraction
It is probably not worth the complexity currently but considering they are using small local CPU models for OCR like tesseract, if they add the support of reading files on the web then I wouldn't be so sure of the CPU bound aspect.
ismailmaj··on The leading AI models are now good historians
20% is a harsh figure but it could be a good entry point to figure out the unknown unknowns and go in depth once you have the relevant keywords using more reliable sources.
ismailmaj··on Every System is a Log: Avoiding coordination in distributed applications
Thanks for the link, not much familiar with CockroachDB but I worked on a similar system that wanted to provide strong consistency when reading from many databases, some of which where Spanner but not all of them.

Eventually the system implemented something similar to "find transaction timestamp at execution time" + optimistic locking, as the article states you lose global linearizability and have to do some re-reads for writes.

What I missed the most from Spanner compared to the system above is that its tricky to do client-side transaction chaining where there are many clients and the order is important.

ismailmaj··on Every System is a Log: Avoiding coordination in distributed applications
Right but it's not magic as you cannot predict networking delays.

The max clock skew also has to pick between A and C, in underestimates you lose serializability, in overestimates you pay in write latency.

ismailmaj··on Every System is a Log: Avoiding coordination in distributed applications
In case of mutation either the replicas will be out of sync (so no C), or you'll need to synchronously mutate the replicas (so bad A due to latency).
ismailmaj··on Every System is a Log: Avoiding coordination in distributed applications
CRDT is really only useful when inconsistencies could be acceptable in some situations and so it depends on the application.

For something that is trying to solve the general problem of consistent single ground-truth log, you can't really do much better than Spanner.

ismailmaj··on LoRA vs. Full Fine-Tuning: An Illusion of Equivalence
How does it compare to partially fine-tuning the model by freezing most of the network beside the last few layers?
← PreviousPage 4 of 5Next →