HNHacker News
TopNewBestAskShowJobs

machiaweliczny

695 karma · joined July 12, 2015

submissionscomments
machiaweliczny··on Europe is in danger of regulating its tech market out of existence
Will have, if Russia can have their own search engine or Czechia (8M people). It’s matter of nit having US competition that undercuts on ads monopoly
machiaweliczny··on AI solves International Math Olympiad problems at silver medal level
This, I've said it many years ago.

Math => Code => Simulation => Robots => GG

machiaweliczny··on OpenAI Announces SearchGPT
The elephant in the room is monetization for people writing this content that perplexity (and then other search engines) will show.
machiaweliczny··on AI solves International Math Olympiad problems at silver medal level
Good, now use DiffusER on algebra somehow please
machiaweliczny··on Node.js adds experimental support for TypeScript
But in practice is was crap
machiaweliczny··on Large Enough
But billionaires already have that, they want a chance of getting their own god.
machiaweliczny··on Large Enough
I tried and sonnet 3.5 can copy most of common UIs
machiaweliczny··on Large Enough
My uneducated guess is that with many layers you can implement something akin to graph in brain by nulling lots of previous later outputs. I actually suspect that current models aren’t optimal with layers all of the same size but i know shit
machiaweliczny··on Large Enough
Just me coding 30% faster is worth it
machiaweliczny··on Large Enough
But projects are great in Sonnet, you just dump db schema some core file and you can figure stuff out quickly. I guess Aider is similar but i was lacking good history of chats and changes
machiaweliczny··on OpenAI slashes the cost of using its AI with a "mini" model
I think they just want to maintain lead with data/task collections and that’s why having cheapest model wins
machiaweliczny··on Show HN: R2R V2 – A open source RAG engine with prod features
Try sonnet 3.5 image understanding.
machiaweliczny··on Waymo One is now open to everyone in San Francisco
Before Uber taxis were basically unaffordable for me. You had to call them and wait in queue till somebody send you one. Check if they aren't cheating etc. I think now much more people are using them due to convenience. Robotaxis can possibly expand market similar how Uber have done it.
machiaweliczny··on Getting 50% (SoTA) on Arc-AGI with GPT-4o
The task provides json grid of numbers
machiaweliczny··on Getting 50% (SoTA) on Arc-AGI with GPT-4o
This challenge looks quite solvable but it's relies on physics understanding and it's has a lot of human/world priors in sense of space understanding and object boundaries.

Seems like it relies on identification of objects and then mapping them somehow. Most of the cases so far that I've seen are based on some transformation or relation between the objects.

So far it seems like some search among common transformatiosn and relations could solve it. Plus some heuristics/computation for counting order, wholeness(boundary) or pattern.

IMO it can be solved by search of programs that combine these + some LLM to guide heuristics most likely.

The only hard one was applied noise or one testing understanding of "gravity".

Did anyone test human baseline for this?

machiaweliczny··on Are animals conscious? New research
Yeah spiders are uniquely smart for their size. I had one using stone and some complicated physics to catch pray (don’t remember exactly the details but was impressed)
machiaweliczny··on Getting 50% (SoTA) on Arc-AGI with GPT-4o
Do you accept such solutions as legit? It’s obviously is easier to generate program that to make prompt that will solve it
machiaweliczny··on Ask HN: If you are starting in 2024, what is the most productive solo dev stack?
What ORM / framework are you using?
machiaweliczny··on Diffusion on syntax trees for program synthesis
I will have to dig into this paper as it looks like exactly this. I wonder if they use closures to limit valid operations space.

The only thing I didn’t understand to make it happen was how to connect it well to description of desired program or edit.

BTW my idea was to train it by destructing programs available on github (so adding noise via some random valid ops and then removing it to retrieve original program). Probably best done in N-1 commit is treated as noise and moving back to commit 0

machiaweliczny··on Diffusion on syntax trees for program synthesis
I had idea about doing something similar based on DifussER paper. One would need to model code edits as algebra similar to add char, replace char or delete char but something like define func, remove func, define var etc. I am undereducated to do it myself but have a feeling it could work.
machiaweliczny··on Google just updated its algorithm. The Internet will never be the same
How about social features like sharing bookmarks or having curated list for specific subjects

When I want to share bookmark folder for other collegues I have to go trough hoops (at least on chrome)

machiaweliczny··on OpenAI plans to announce Google search competitor on Monday
I hoped that google would add a way to chat on websites with likeminded people in their browser and to merge it somehow with groups and search history.

It's more organic than tweeter/X and something that should be built-in part of web IMO.

The problem is probably that without accounts you can't fight with bots easily

machiaweliczny··on Bicycle use now exceeds car use in Paris
In NYC you should also have some law that prohibit's going more than 20mph on bike, as some e-bikers are a bit crazy.
machiaweliczny··on 'Irresponsible' to ignore consciousness across animal world scientists argue
Yeah I uses to have these but currently don’t
machiaweliczny··on Llama 3 8B is almost as good as Wizard 2 8x22B
There's still lot's of low hanging fruit in data preparation for these models it seems.

I wonder if we will reach insane quality when we will be able to allow models to generate not exact same text but something similar (thus allowing for more compression). I think math leans a lot to this type of training as one can super easily verify if generated result is fine or not.

We probably should do similar thing for language when other model judges if generation is ok or not (but doesn't require exact generation). This probably can be done as fine tuning? Anyone know if this has been tested already?

machiaweliczny··on HonoJS: Small, simple, and ultrafast web framework for the Edges
Do you have some example?
machiaweliczny··on Show HN: Next-token prediction in JavaScript
Man watch video from Karpathy and learn about prefix trees (tries)
machiaweliczny··on PHP in 2024
Laravel is Rails clone afaik
machiaweliczny··on Grok
If they are so behind they could make it open source instead of open weights and get some help.
machiaweliczny··on Quiet-STaR: Language Models Can Teach Themselves to Think Before Speaking
It’s a bit over my head for now but seems like GFlowNets are tackling this problem a bit.
← PreviousPage 6 of 29Next →