jphoward

The browser it built, obviously the context window of the entire project is huge. They mention loads of parallel agents in the blog post, so I guess each agent is given a module to work on, and some tests? And then a 'manager' agent plugs this in without reading the code? Otherwise I can't see how, even with ChatGPT 5.2/Gemini 3, you could do this otherwise? In retrospect it seems an obvious approach and akin to how humans work in teams, but it's still interesting.

Here is some more information: https://en.wikipedia.org/wiki/All_Red_Line

And one of the old cable huts still exists: https://en.wikipedia.org/wiki/Pacific_Cable_Station

'Ruined' is a bit strong, isn't it?

It seems to be like some of the scales slightly off?

If you are looking at the ladybird (ladybug) with the amoeba to the left, the amoeba isn't an order of the magnitude smaller - it would actually be visible by the human eye (bigger than a grain of sand)? Indeed, the amoeba seems the same size as the ladybird's foot?

Similarly, this makes the bumblebee appear smaller than a human finger (the in the adjacent picture), which isn't the case?

Nice! What might be a nice lesser 'clue' to simply revealing a word is highlighting letter(s) on the board that are part of it? Favouring maybe highlighting letters that are contiguous with a blue bit?

Have you not seen some of the replies at the link?

For example:

"You are joking ?!

The commit about source only is 4 days old (9e49d5e)

We are currently paying for a license while using the open source version, you already removed the oidc code from UI console and now docker images. We are not happy by this lock-in. We will discuss this internally, but you may loose a paying customer with this behavior."

Are you me? Exact same! The problem with dual monitors is either you're sat in front of the gap, or you need to pivot. This way you get a 'normal' monitor and a portrait section to the side, much better.

Why does that link say the current director of public prosecutions is Sir Kier Starmer? It's hard to take it seriously.

$49 seems a surprisingly high amount for something aimed at students and learners - I appreciate the content may be good, but it's effectively 3 times a Netflix subscription.

It's meant to be something you stick with in the "long term" by its nature, and yet an annual subscription is $500 - this is just completely unrealistic for any student. Someone in a lower end job hoping to "up skill" is going to really struggle with this.

I don't think there's any question about the "probably" - I'm not that involved in the community but the subreddit makes it sound totally disharmonious.

"There are far fewer East Asian CEOs in the Fortune 500, and most of them are the founders of their companies like Jensen Huang (Nvidia), Tony Xu (DoorDash), Lisa Su (AMD)"

Lisa Su only joined AMD in 2012?

I think they could release non-agentic models that are as good as 4o, and have almost no repercussions on sales tbh.

I have Ollama installed (only a small proportion of their clients would have a large enough GPU for this) and have download deepseek and played with it, but I still pay for an OpenAI subscription because I want the speed of a hosted model, and never mind the luxuries of things like Codex's diffs/pull request support, agents on new models, deep research etc. - I use them all at least weekly.

Yeah but because it's a GitHub repo is has an inherent audit trail for that, so it's not really erasing misdeeds... indeed it highlights those people in diffs!

At the bottom it references a GitHub where people have previously added signatures against Jon Pretty - and now the maintainer says "NOTE: This repo is closed. Do not open issues; they will be summarily closed and ignored." - i.e. telling people they shouldn't even TRY to amend their signatures.

Regardless of what you think of Jon Pretty, how is this justifiable? Telling people they can't unsupport something because you're not open to issues, but also not removing it?!

64 GB of RAM which is addressable by a GPU is exceptional for a laptop - this is not just system RAM.

No but 64 GB of unified memory provides almost as much GPU RAM capacity as two RTX 5090s (only less due to the unified nature) - top of the range GPUs - so it's a truly exceptional laptop in this regard.

Sorry I'm a bit out of the loop, but what's CapCut?

I don't think you know what "cringe" means... this really makes you cringe? And they didn't say it was crazy, they said it was interesting enough to raise eyebrows. Everyone knows there are amazing coders in hedge funds, but not many hedge funds have forked a language - it is worthy of discussion here.

Are you sure? Philips' medical side have an in house GUI library (Qt-esuqe) called Sense and it's mainly implemented using C++.

I don’t understand how this works with systems like toothpaste? I get if you somehow stored your toothpaste in some queue-like structure where you only see 1 at a time, then yes, reaching the last would be helpful. But surely they are in a cupboard or drawer or something?

To know to start the sentinel toothpaste he must search around and prove it’s the only one left - by doing this he doesn’t need the sentinel anyway, as the searching process with yield only one, sentinel or not?

If it doesn’t need a search, e.g. they are in a line/all on easy view - then it’s obvious it’s the last even without a sentinel?

Previously discussed at https://news.ycombinator.com/item?id=26967283

Is anyone is the AI/ML area finding success with anything other than conda, where installation of CUDA/CUDnn is required? Although I often have to pip install a lot of packages, I find conda's nvidia/pytorch/conda-forge channels are still by far the easiest way to get a deep learning stack up and running, and so I just stick with conda environments. I've tried poetry in the past but getting the NVidia deep learning stack up and running was really tough.

I was excited about JAX, but I think the developers missed a trick when they decided it should be entirely immutable. It sounds silly, but I think if I have an array `x` and want to set index 0 to 10, it's a big mistake if I can't do:

  x[0] = 10

And instead I have to do:

  y = x.at[0].set(10)

Of course this has advantages, and I know it sounds lame, but as someone whose brain works in numpy, this is really offputting.

I like opencv, but can there be a stretch goal where if they raise $1M they agree to move from Blue/Green/Red colour channel order by default to RGB in OpenCV5, like everyone else on planet earth? I genuinely avoid opencv for this reason because I know I will forget to BGR->RGB on loading and RGB->BGR before saving occasionally, and so I'd rather stick to skimage/imageio/PIL. Or hell, add an `as_rgb=False` parameter to the load/save functions or something!

The problem is the discussed results are comparing proportions of a relatively small number - 67 questions. If you model this as a binomial distribution, then 62/67 which GPT4-turbo got gives a 95% confidence interval of the 'true' performance of 83.4% to 97.5%, ie it comfortably includes the proportion that GPT4 achieved (64/67=95.5%).

I think the evidence from these tests are not strong enough to draw conclusions from.

One thing I always wonder is how did all these countries manage to find enough gold to run an (albeit tiny) economy off them? I've never heard of/seen a gold mine in the UK, and yet 2000 years ago they were mining enough to mint currency. Was it all relatively surface level and rapidly mined out, and now all gone?

I think JAX is cool, but I do find it slightly disingenuous when it claims to be "numpy by on the GPU" (as opposed to PyTorch), when actually there's a fundamental difference; it's functional. So if I have an array `x` and want to set index 0 to 10, I can't do:

  x[0] = 10

Instead I have to do:

  y = x.at[0].set(10)

Of course this has advantages, but you can't then go and claim that JAX is a drop in replacement for numpy, because this such a fundamental change to how numpy developers think (and in this regard, PyTorch is closer to numpy than JAX).

You can see regularly in practice where aggressive data augmentation is used, which obviously is only used on training data. But, of course, you'd still be 'overfit' if you fed in unaugmented training data.

It's a problem when it happens suddenly because you end up with a 'top heavy' age distribution. The elderly pensioners:young earners ratio explodes and society security/national insurance cannot afford healthcare/pensions etc.

This is compounded by what was (until recently) a steady increase in life expectancy.

The really frustrating aspect of this is pensioners, who have traditionally huge voter turnouts during elections, are unlikely to have their benefits cut, because it would be political suicide for a party. The elderly believe they have "paid in" to their plans and deserve them, but outside of private pensions, this is usually a significant over-simplification. Instead, what will happen is the working young will need to pay for the generous promised benefits that the elderly are receiving, whilst 'paying in' for much lower benefits for themselves.

I love Nim so much. I really hope some big company decides to use it as their workhorse language - it feels like the only thing holding it back is the lack of corporate 'buy in'. Most similarly good languages (except F# maybe?) seem to get lucky with becoming a poster boy for at least one company.