Hacker News: dandelionv1bes

New comment by dandelionv1bes in "Understanding Neural Network, Visually"

dandelionv1bes — Sun, 08 Feb 2026 09:32:43 +0000

If you’re interested in an easy explainer for backprop, then highly recommend math for deep learning by Kneusel. Finished it recently and you can apply the chain rule backwards by hand with a tiny NN, and code one too.

New comment by dandelionv1bes in "The $100B megadeal between OpenAI and Nvidia is on ice"

dandelionv1bes — Sat, 31 Jan 2026 09:09:06 +0000

Saw Empire of AI in a bookshop recently but held off buying as wasn’t sure if it was going to be surface level. You’d recommend?

AI-exposed jobs deteriorated before ChatGPT

dandelionv1bes — Fri, 23 Jan 2026 08:56:52 +0000

Article URL: https://arxiv.org/abs/2601.02554

Comments URL: https://news.ycombinator.com/item?id=46730164

Points: 2

# Comments: 0

AI–AI bias: LLMs favor communications generated by large language models

dandelionv1bes — Wed, 21 Jan 2026 19:41:00 +0000

Article URL: https://www.pnas.org/doi/10.1073/pnas.2415697122

Comments URL: https://news.ycombinator.com/item?id=46710511

Points: 2

# Comments: 0

New comment by dandelionv1bes in "Targeted Bets: An alternative approach to the job hunt"

dandelionv1bes — Tue, 20 Jan 2026 10:57:38 +0000

I struggle with this a bit because while my network isn’t bad - I really can’t stand social network like threads/X etc - I’m not on any social media bar here and LinkedIn.

Do you think investing in bluesky is worth it? I’m in industry but have a PhD ongoing in TTI models so I should probably get on it :/

New comment by dandelionv1bes in "Predicting OpenAI's ad strategy"

dandelionv1bes — Sun, 18 Jan 2026 14:46:04 +0000

Nice read - I can’t seem to subscribe? Getting an error with Buttondown.

Show HN: Tiny Toy Network – a neural net to practice backpropagation

dandelionv1bes — Sun, 18 Jan 2026 14:33:50 +0000

I’ve recently finished Math for Deep Learning and really enjoyed computing back propagation by hand. However, it’s one of those things where if you stop doing it for six months, you’ll just forget how (at least for me!). I put together this tiny neural network based on the example from Kneusel so I can practice my math regularly. Welcome any feedback / bug spotting, or similar tools to help with practicing these skills.

Comments URL: https://news.ycombinator.com/item?id=46668079

Points: 1

# Comments: 0

New comment by dandelionv1bes in "ThinkNext Design"

dandelionv1bes — Sun, 18 Jan 2026 09:20:27 +0000

i’m sorry this really made me laugh - I hope she’s ok!

New comment by dandelionv1bes in "ClickHouse acquires Langfuse"

dandelionv1bes — Sat, 17 Jan 2026 10:58:10 +0000

I think the docs are great to read, but implementing was a completely different story for me, ie, the Ask AI recommended solution for implementing Claude just didn’t work for me.

They do have GitHub discussions where you can raise things, but I also encountered some issues with installation that just made me want to roll the dice on another provider.

They do have a new release coming in a few weeks so I’ll try it again then for sure.

Edit: I think I’m coming across as negative and do want to recommend that it is worth trying out langfuse for sure if you’re looking at observability!

New comment by dandelionv1bes in "ClickHouse acquires Langfuse"

dandelionv1bes — Sat, 17 Jan 2026 10:44:45 +0000

I do understand why it’s a product - it feels a bit like what databricks has with model artifacts. Ie having a repo of prompts so you can track performance changes against is good. Especially if say you have users other than engineers touching them (ie product manager wants to AB).

Having said that, I struggled a lot with actually implementing langfuse due to numerous bugs/confusing AI driven documentation. So I’m amazed that it’s being bought to be really frank. I was just on the free version in order to look at it and make a broader recommendation, I wasn’t particularly impressed. Mileage may vary though, perhaps it’s a me issue.

New comment by dandelionv1bes in "Ask HN: What did you find out or explore today?"

dandelionv1bes — Thu, 15 Jan 2026 09:03:24 +0000

I’ve been slowing crunching through Math for Deep Learning, so spent a fair amount of time looking at Hessian matrices + second order optimisation. I’ve been slowly reading this book for a year, so stopping to do most of the math by hand each time. One chapter to go!

Then I was sick all last week, so ended up down a rabbit hole about the current card collecting bubble (right word?). Super interesting.

Show HN: Llama 4 Maverick explores Japantown in Watch Dogs 2

dandelionv1bes — Sun, 11 Jan 2026 14:39:06 +0000

This is a little project we’ve been working on, which started as multimodal models physically controlling gamepads. It’s evolved a little bit since then and we now have a setup where an agent can control movement, cameras and taking selfies (where game allows) in an open world.

The project sprung actually from wanting to try out some observability platforms for LLMs, and sort of kept evolving. Hope you enjoy this little free roam around SF. It’s helped us learn a lot about prompting, observability and actuation of agents in the process.

We’ll be posting more hopefully, but very much a “hack” as it won’t take any time away from us actually playing these titles. Hope you enjoy this video and encourage you to play this awesome game.

Comments URL: https://news.ycombinator.com/item?id=46576120

Points: 2

# Comments: 0

AgentRoam: Watch GPT-5.2 control movement, camera and selfies in Watch Dogs 2

dandelionv1bes — Sat, 10 Jan 2026 20:19:39 +0000

Article URL: https://www.youtube.com/watch?v=XTYWewHz-Tg

Comments URL: https://news.ycombinator.com/item?id=46569532

Points: 2

# Comments: 0

New comment by dandelionv1bes in "Robotopia: A 3D, first-person, talking simulator"

dandelionv1bes — Sat, 10 Jan 2026 10:48:56 +0000

Amazing! Thanks will join.

New comment by dandelionv1bes in "Robotopia: A 3D, first-person, talking simulator"

dandelionv1bes — Sat, 10 Jan 2026 10:07:04 +0000

This is fantastic. I think it’s nailed in the substack what was missing from a lot of these LLM driven NPCs that did not feel authentic. I have a couple of follow-up questions on specifics relating to analysis of behaviour with LLMs (in game-dev myself). Would it be possible to speak to you directly on them?

New comment by dandelionv1bes in "Favorite Tech Museums"

dandelionv1bes — Sat, 10 Jan 2026 10:02:22 +0000

GAMM in Rome is quite fun (Italy):

https://www.gammuseum.com/en/

Shout-out to the museum of Ancient Greek technology, with their wine automaton (Athens, Greece): https://kotsanas.com/

New comment by dandelionv1bes in "Favorite Tech Museums"

dandelionv1bes — Sat, 10 Jan 2026 10:00:22 +0000

Bletchley is great! I need to go again

New comment by dandelionv1bes in "Strange.website"

dandelionv1bes — Tue, 06 Jan 2026 08:57:15 +0000

Haven’t seen Wiby before but it’s fantastic, thank you.

New comment by dandelionv1bes in "Why didn't AI “join the workforce” in 2025?"

dandelionv1bes — Mon, 05 Jan 2026 22:23:45 +0000

The response to the Sal Khan op-ed resonated with me, along with other parts of this article. Something I’ve been digging more into is some of the figures around proposed job losses from AI. I think I even posted a simulation paper last week.

After posting that, I came across numerous papers which critique Frey & Osborne’s approach, who are some of the forefathers for the AI job losses figures we see banded around commonly these days. One such paper is here but i can dig out others: https://melbourneinstitute.unimelb.edu.au/__data/assets/pdf_...

It has made me very cautious around bold statements on AI - and I was already at the cautious end.

New comment by dandelionv1bes in "Agentic Patterns"

dandelionv1bes — Mon, 05 Jan 2026 22:13:56 +0000

Thank you for putting it together. I looked at a couple of the references and they look like they point to your blog. Do you have a view at all of popular patterns in terms of citations? Might be useful