frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

Mercury: Ultra-Fast Language Models Based on Diffusion

https://arxiv.org/abs/2506.17298
184•PaulHoule•3h ago•67 comments

I used o3 to profile myself from my saved Pocket links

https://noperator.dev/posts/o3-pocket-profile/
117•noperator•3h ago•58 comments

Launch HN: Morph (YC S23) – Apply AI code edits at 4,500 tokens/sec

34•bhaktatejas922•1h ago•14 comments

Adding a feature because ChatGPT incorrectly thinks it exists

https://www.holovaty.com/writing/chatgpt-fake-feature/
128•adrianh•1h ago•31 comments

When Figma starts designing us

https://designsystems.international/ideas/when-figma-starts-designing-us/
110•bravomartin•1d ago•44 comments

François Chollet: The Arc Prize and How We Get to AGI [video]

https://www.youtube.com/watch?v=5QcCeSsNRks
102•sandslash•4d ago•77 comments

Dyson, techno-centric design and social consumption

https://2earth.github.io/website/20250707.html
19•2earth•1h ago•7 comments

SUS Lang: The SUS Hardware Description Language

https://sus-lang.org/
3•nateb2022•6m ago•0 comments

Solving Wordle with uv's dependency resolver

https://mildbyte.xyz/blog/solving-wordle-with-uv-dependency-resolver/
52•mildbyte•1d ago•4 comments

Bitchat – A decentralized messaging app that works over Bluetooth mesh networks

https://github.com/jackjackbits/bitchat
548•ananddtyagi•16h ago•245 comments

Tuning the Prusa Core One

https://arachnoid.com/3D_Printing_Prusa_Core_One/
21•lutusp•1h ago•10 comments

Hymn to Babylon, missing for a millennium, has been discovered

https://phys.org/news/2025-07-hymn-babylon-millennium.html
96•wglb•3d ago•22 comments

Lightfastness Testing of Colored Pencils

https://sarahrenaeclark.com/lightfast-testing-pencils/
31•picture•2d ago•4 comments

Cpparinfer: A C++23 implementation of the parinfer algorithm

https://gitlab.com/w0utert/cpparinfer
33•tosh•4d ago•2 comments

The Era of Exploration

https://yidingjiang.github.io/blog/post/exploration/
10•jxmorris12•58m ago•1 comments

Neanderthals operated prehistoric “fat factory” on German lakeshore

https://archaeologymag.com/2025/07/neanderthals-operated-fat-factory-125000-years-ago/
175•hilux•3d ago•116 comments

A non-anthropomorphized view of LLMs

http://addxorrol.blogspot.com/2025/07/a-non-anthropomorphized-view-of-llms.html
321•zdw•17h ago•300 comments

Show HN: NYC Subway Simulator and Route Designer

https://buildmytransit.nyc
49•HeavenFox•2h ago•3 comments

Show HN: I wrote a "web OS" based on the Apple Lisa's UI, with 1-bit graphics

https://alpha.lisagui.com/
430•ayaros•21h ago•124 comments

Show HN: Piano Trainer – Learn piano scales, chords and more using MIDI

https://github.com/ZaneH/piano-trainer
141•FinalDestiny•2d ago•42 comments

Showh HN: Microjax – JAX in two classes and six functions

https://github.com/joelburget/microjax
27•joelburget•2h ago•1 comments

CPU-X: CPU-Z for Linux

https://thetumultuousunicornofdarkness.github.io/CPU-X/
5•nateb2022•2h ago•1 comments

Anthropic cut up millions of used books, and downloaded 7M pirated ones – judge

https://www.businessinsider.com/anthropic-cut-pirated-millions-used-books-train-claude-copyright-2025-6
171•pyman•7h ago•201 comments

The Cat's Meat Man: Feeding Felines in Victorian London

https://publicdomainreview.org/essay/the-cats-meat-man/
52•ohjeez•2d ago•6 comments

Intel's Lion Cove P-Core and Gaming Workloads

https://chipsandcheese.com/p/intels-lion-cove-p-core-and-gaming
231•zdw•17h ago•84 comments

Why English doesn't use accents

https://www.deadlanguagesociety.com/p/why-english-doesnt-use-accents
243•sandbach•19h ago•360 comments

Building the Rust Compiler with GCC

https://fractalfir.github.io/generated_html/cg_gcc_bootstrap.html
200•todsacerdoti•18h ago•51 comments

AI Cameras Change Driver Behavior at Intersections

https://spectrum.ieee.org/ai-intersection-monitoring
6•sohkamyung•4h ago•0 comments

LLMs should not replace therapists

https://arxiv.org/abs/2504.18412
209•layer8•18h ago•275 comments

The Harvey Edwards Archive

https://www.harveyedwards-archive.com
3•toomuchtodo•22m ago•0 comments
Open in hackernews

Poland's clean energy usage overtakes coal for first time

https://www.ft.com/content/ae920241-597e-49d9-a4b9-bfdfa9deabb6
86•stared•7h ago

Comments

stared•7h ago
No paywall: https://archive.is/5A0kr

The original analysis of renewables in Poland in June 2025: https://www.forum-energii.eu/en/monthly-magazine-1

baranul•6h ago
Meanwhile in America, oil and coal is becoming king again. Climate change you say? Oh well, "thoughts and prayers".
watwut•5h ago
Why is this downvoted? It is an accurate observation of the politics.
viraptor•5h ago
Because it's shallow and ragebaity. If anyone wants to discuss that topic, there's so many quality articles these days to submit and talk about.
atwrk•5h ago
Edit: I snarkily replied to an imagined comment, not the one that was actually written...
dewey•5h ago
The parent post says America, not Africa though.
atwrk•5h ago
Oh wow, I stand corrected. No idea how I could have misread that.
passwordoops•5h ago
The typical answer to Climate Change isn't "thoughts and prayers" but "go fck yourself"
bryanlarsen•5h ago
Even America is building far more renewables than fossil fuel or nuclear, despite the deck being stacked against them.
MarcelOlsz•4h ago
How many days worth of coal burning does 90 days of billionaire jets travelling to Bezos wedding count as?
hwillis•2h ago
A 747 burns ~9 tonnes of kerosene per hour, creating ~29.6 tonnes CO2. The Monroe Power Plant produces 3400 MWe at ~1 kg CO2 per kWh, so ~3400 tonnes CO2 per hour.

It's a little complicated to weigh stratospheric emissions- the CO2 has a larger impact, and while the water droplets and contrails left by planes somewhat counteracts it (by reflecting incoming infrared) it's harder to compare intangibles like mercury emissions from coal. If you just say its all a wash, that plant is equivalent to 120 747s running full speed.

Private jets consume more like .9-1.5 tonnes per hour, so that's equivalent to ~900 billionaires. That's a bit less than half of them which is probably a lot more than were at the wedding. They also probably parked them instead of leaving them circling in the air.

If there were 90 billionaires who flew 12 hours each way in their private jets, then they probably released around 2.5 hours worth of Monroe Power Plant time over those 90 days- 8458 tonnes. Fun fact, the pilots and flight attendants probably used ~1000 tonnes of CO2 worth of energy etc and exhaled ~25 tonnes of CO2 in that time. 25 tonnes is small compared to the planes (>.3%), but in those 90 days the planes released just .11% as much as the coal plant.

Coal really truly sucks and it's unfair that I can't eat tuna without getting mad hatter disease.

infecto•4h ago
As much as I don’t like the administration, I don’t think this is necessarily true. Subsidies are being removed but I think most renewable has the chance to hold on its own. I do wish more of the markets worked closer to a an actual market like Texas which incentivizes creativity and trying to maximize. Renewables are so cost effective these days that most areas implement them and then natural gas serves as a useful baseload.
krige•4h ago
At the same time the admin has undercut and gutted many climate tracking projects [0] so it's really not hard to see the underlying intentions.

[0] for instance https://apnews.com/article/climate-change-national-assessmen...

adrianN•6h ago
Renewables are just really cheap. Unless the government is actively harming them for ideological reasons they get built. Baseload power plants have a harder time making a profit every year.
Xelbair•5h ago
Are they? Usually most analysis ignore costs of having a buffer in the system - either batteries or water reservoirs for pumped storage. That also inclides efficiency of storage, and the fact that cycles of power generation in most cases do not align with usage.

Initial cost per unit is low and they're faster to build - that's the reason of their proliferation - and in case of personal use, subsidiaries.

Compared to nuclear - which takes many political terms to build - politicians can reap benefits early. Because nuclear is superior by every metric except:

- high initial costs

- longer lead time

Not to mention less land required, which is another of ignored costs.

bryanlarsen•4h ago
It's not just high initial costs, nuclear also has significant running costs, high disposal costs and massive time costs.
Y-bar•4h ago
> significant running costs

Exactly this.

The running costs per produced MW is so high that governments has to promise to effectively pay billions in subsidies to NPP:s both in terms of cheap state-backed loans and contract-for-difference power price because because they would otherwise not be viable.

[1] https://www.reuters.com/business/energy/sweden-proposes-stat...

mpweiher•4h ago
That's actually not true. Running costs of nuclear power plants are extremely low, some of the lowest of all energy sources.

The reason that NPP operators now require guarantees from governments is to offset the incredibly lopsided benefits offered to intermittent renewables, including subsidies, cheap loans, guaranteed income regardless of demand and priority.

Nobody can effectively compete with the government handing out those kinds of benefits.

Y-bar•4h ago
If running costs are so low as you claim, why does both the Swedish Government as well as plant owners such as Fortum claim that the only way to be profitable on the market is for the Government to subsidise the cost of running the plants?

Wind or Hydro in Sweden gets no such unique subsidies.

p_l•4h ago
Because markets, as currently implemented for energy in EU, do not adequately include things like intermittency of demand.

If you take a source that has very low per-MWh price but very intermittent, it gets effectively highly prioritized and everyone else is paying price to match up with this wrecking ball. Except, usually, gas turbines which slot very very nicely into the swings in capacity related to wind and solar and thus benefit from them.

bryanlarsen•3h ago
That's true for the UK and Spain as of 2023[1], but the other major energy markets in the EU incorporate significant capacity mechanisms in their pricing.

Edit: Spain added one in Feb 2025. Haven't looked up the UK yet.

1: https://www.eurelectric.org/in-detail/capacity-mechanisms/

Y-bar•3h ago
That's not true in Scandinavia (can't speak for EU as a whole). For example Nord Pool's market structure and pricing mechanisms are specifically designed to account for the intermittency of production and demand: https://www.nordpoolgroup.com/en/the-power-market/Intraday-m...
bryanlarsen•4h ago
Those are generally available to anybody providing either carbon neutral power or grid stability. Nuclear can take advantage of those too.

Nuclear generally receives other subsidies too -- like cost overrun protection, insurance, and cleanup waivers.

ViewTrick1002•2h ago
Running costs are low but not extremely low.

They are vastly undercut by renewables and which is now starting to dig into their capacity factors.

Nuclear power simply is the worst possible fit for modern grids.

soco•9m ago
"Nuclear power simply is the worst possible fit for modern grids." because it's too stable? Or what do you mean?
js8•4h ago
Actually, not quite true, according to this:

https://www.cleanenergywire.org/news/large-solar-arrays-batt...

Nuclear also has non-negligible ecological impact outside the power plant due to mining and processing of nuclear fuel.

hwillis•3h ago
> and the fact that cycles of power generation in most cases do not align with usage.

This is false. Power usage everywhere is highest when the sun is shining. There are also very very few places where solar power ever causes the market to bottom out with any regularity. Note that there is no technical problem with this- you can always just disconnect renewables from the grid.

The phenomenon you are thinking of -the duck curve- refers to the power demand after subtracting solar. The daily peak consumption of power in many places is wider than solar generation, so if there is enough solar you end up getting new smaller peaks just after dawn and around sunset. This is minorly inconvenient for non-renewable sources, which prefer to have more predictable demands.

> Usually most analysis ignore costs of having a buffer in the system

Correctly! Non-renewable plants are the ones that need buffer. Solar, wind, hydro etc can all be connected to a grid with zero instability- you just unplug them if nobody wants the power. Non-renewable plants have slow ramp speeds- they need the buffer in order to follow a changing load.

> Not to mention less land required, which is another of ignored costs.

This is incorrect; I don't know of any analyses which don't include land and interconnection costs which are obviously substantial. If you mean more intangibly... that's very silly. The US Interstate system is 3.9 million miles of road, with 60' medians, 16' of shoulder, and 48' of lanes. 237,250 square kilometers. The "blue square"[1] is 10,000 square km. The amount of land we spend on parking lots absolutely dwarfs it.

> Because nuclear is superior by every metric

Nuclear has not gotten cheaper- why would it? It's a big clockwork. We are not better at building pipes than we were 80 years ago. Solar has and will continue to: plants get more productive, panels get thinner, efficiencies go up. There is no grounding principle that indicates nuclear can be cheaper, and it certainly is not in practice. Solar is far cheaper than coal by capacity much less kWh, and nuclear plants are more complex than coal. What indicates that a 500 MW nuclear plant should be cheaper than a 500 MW coal plant, not counting running costs?

> Are they?

Demonstrably yes, absent weird conspiracy theories. Renewable installations keep opening at much lower costs than traditional plants.

[1]: https://blogs.ucl.ac.uk/energy/2015/05/21/fact-checking-elon...

sofixa•2h ago
> This is false. Power usage everywhere is highest when the sun is shining

While the biggest peak is around midday, the second biggest is in the evening (most people are home, cooking, watching TV/listening to music/playing video games/etc; or in restaurants, clubs, cinemas, etc) which, depending on location and time of year, can easily be after sunset (e.g. half the year in the Northern hemisphere for sure). You still need enough power to cover that, especially if it has been a cloudy/rainy day, or week, or month.

> Non-renewable plants are the ones that need buffer. Solar, wind, hydro etc can all be connected to a grid with zero instability- you just unplug them if nobody wants the power. Non-renewable plants have slow ramp speeds- they need the buffer in order to follow a changing load

And this is so easy and foolproof to do, just check out the Iberian power outage.

hwillis•2h ago
> You still need enough power to cover that, especially if it has been a cloudy/rainy day, or week, or month.

That's besides the point! The window of highest demand completely covers the window of solar. You can build a LOT of solar before storage starts becoming cheaper than just building more solar. You only need storage if it's ALL solar- you can have a majority of your power supplied by solar with hardly any storage! There this idea that if you overbuild solar that power will have nowhere to go, or something- you can just turn it off. You use backup power for the non-shining hours and you're totally fine.

> And this is so easy and foolproof to do, just check out the Iberian power outage.

In fact I did[1]. Page 117: "In fact, in most of the network nodes analyzed, there is no correlation between voltage stability and the amount of solar generation or the amount of coupled synchronous generation"

They had 2.3 seconds of inertia, more than the regulated 2 seconds. Power sloshing through interconnects caused plant ramp rates to be overwhelmed one-by-one, causing the cascading failure, because they had no buffering. If they were all solar or wind plants, the failure would not have happened!

[1]: https://media.licdn.com/dms/document/media/v2/D4D1FAQGcyyYYr...

rich_sasha•2h ago
> > and the fact that cycles of power generation in most cases do not align with usage.

> This is false. Power usage everywhere is highest when the sun is shining.

This is kind of true but also not. In Poland, and in a lot of Europe, power usage is by far the highest in winter, day or night, for residential heating. That's also the time with little sun: days are short and the sky tends to be heavily overcast. Sunny weather is rare in winter.

hwillis•2h ago
Numbers: In poland the solar output in December is 1/5th of July output while demand is ~10% higher. Insolation drops off quickly at higher latitudes so it dominates much more strongly than heating demand even if heating were electrical.

At the global scale this would ideally balance out- more equatorial solar is cheaper in the winter since they don't need air conditioning, so you just send it up north. That's the only really feasible solution to seasonal variation in individual countries- it's totally unreasonable to store 3 months worth of power. Its also important to note that even with 2x or 3x oversupply, solar is cheaper than nuclear currently is.

[1]: https://re.jrc.ec.europa.eu/pvg_tools/en/#PVP [2]: https://www.pse.pl/web/pse-eng/data/polish-power-system-oper...

bryanlarsen•1h ago
Overbuilding is significantly cheaper than either months worth of storage or intercontinental HVDC electrical links. It's probably worth building links with Romania, but probably not equatorial North Africa. Instead if you want to hit 100% carbon free, overbuild solar to almost your 5X, mix in a good amount of wind, a couple days of storage and links with Romania or nearby.

IMO 99% is a much better target than 100%: almost all of the benefits for a lot less cost. Idle natural gas plants don't emit carbon.

(And by 100% I mean 99.99%. The grid isn't reliable enough to make more than 99.99% worth the cost).

adrianN•3h ago
You can go to like 60-80% renewable with barely any storage in the system. Few countries are at the stage of the renewable rollout where they need to think about storage. Poland can easily double the amount of renewables without additional storage.
sofixa•2h ago
> You can go to like 60-80% renewable with barely any storage in the system

That depends on the renewable type (e.g. 60% hydro is fine; 60% solar is not, because the sun goes down at night, and there can be extended periods of overcast weather which lower solar production) and what backup sources (like gas peaker plants) are available.

bryanlarsen•2h ago
60-80% can be usually be met with a well designed mix of hydro & solar.

You can get to 100% with either hydro or geothermal, but most countries don't have that option.

adrianN•2h ago
Germany has similar climate to Poland and manages around 60% renewable (and growing) with very little storage. A mix of wind and solar with fossil power plants for backup works reasonably well.
ViewTrick1002•2h ago
Why should I as a consumer buy electricity from your extremely expensive new built nuclear power when either my own renewables with storage or grid based renewables with storage delivers?

Nuclear power is extremely expensive and doesn’t provide anything a modern grid needs.

jillesvangurp•5h ago
Even when governments interfere, people will still invest in grid independence, resilience, or just access to cheaper rates. Domestic solar makes sense only because the grid is unreasonably expensive for a lot of people.

Base load is one of those terms that gets wielded without putting numbers on it. It's kind of meaningless without numbers. How many gw of it is needed? Is it whatever we have? Or far less than that? Considering that most countries have been actively removing lots of base load in the form of coal plants and have seen a lot of growth in renewables, you could make the point that whatever that number is, it's far less than it used to be. For the simple reason that a lot of it disappeared without creating a lot of instability.

Coal plants don't have much of a future. Gas plants are more flexible but seem to be increasingly used for reserve power rather than for base load and of course they compete with batteries for that. And the less they are utilized, the less profitable they get. Neither is attractive from an investment point of view.

vv_•4h ago
> people will still invest in grid independence

Most inverters don't work without grid synchronization. E.g. you lose electricity from your provider and your batteries / stored energy won't work either.

All new projects need to be A++ energy class rated which require you to use renewable energy, which is likely one of the main reasons for these increases.

ViewTrick1002•2h ago
Island mode is a trivial option adding negligible cost when building a home solar and battery system.
Havoc•6h ago
Progress...but they're still consistently the dirtiest country in the region when it comes to energy mix

https://app.electricitymaps.com/map/72h/hourly

viraptor•5h ago
The history of coal mining in Poland is strong and those families still have an impact on what's happening, even with many mines closing. https://www.politico.eu/article/polish-parties-scramble-for-...
stared•5h ago
So it is why these changes matter - d/dt has an opportunity to be very high.