Hacker Newsnew | past | comments | ask | show | jobs | submit | ffsm8's commentslogin

Strange, you shouldn't have issues with vacuums on queue tables unless you're doing it wrong?

Were you not using partitions like this?

CREATE TABLE events_2026_04 PARTITION OF events FOR VALUES FROM ('2026-04-01') TO ('2026-05-01');

CREATE TABLE events_2026_05 PARTITION OF events FOR VALUES FROM ('2026-05-01') TO ('2026-06-01');

https://www.postgresql.org/docs/current/ddl-partitioning.htm...

> Bulk loads and deletes can be accomplished by adding or removing partitions, if the usage pattern is accounted for in the partitioning design. Dropping an individual partition using DROP TABLE, or doing ALTER TABLE DETACH PARTITION, is far faster than a bulk operation. These commands also entirely avoid the VACUUM overhead caused by a bulk DELETE.

It was a lot more annoying earlier then pg 13 though, maybe you're just reminiscing things from the 2010s?


Let's not get reality get in the way of their headfiction of how great Muslim rule is

Usually they're hemorrhaging performance while training.

From that it's pretty likely they were training mythos for the last few weeks, and then distilling it to opus 4.7

Pure speculation of course, but would also explain the sudden performance gains for mythos - and why they're not releasing it to the general public (because it's the undistilled version which is too expensive to run)


Mythos is speculated to have 10 trillion parameters. Almost certainly they were training it for months.

Naturally, it is however noticeable that in the lead up to a model release we always get massively degraded performance for the preceeding few weeks

It's been like that for each model release within the last year


but how true is this? this is almost impossible to measure and those that do[1] find no significant difference

i personally haven't noticed any downgrade at all.

it's entirely possible there's a mass delusion going on where everyone gets wowed by 4.6 initially, then accepts the new baseline and gets used to it, then thinks that baseline is no longer impressive and thus degraded

it doesn't help that anthropic changed defaults for its claude code harness for all users suddenly

the best and only evidence i've seen for actual degradation is that the web version of opus 4.6 failed the car wash test, and since you cannot simply choose to "disable adaptive thinking" and other parameters with the web version, you truly may have gotten a worse product

[1] https://marginlab.ai/trackers/claude-code-historical-perform...


And at that scale (1k) it ain't even that hard, a single room could be enough to hazardly drop them on shelves with a big fan to draw out the heat

You consider it normal working condition if you're right next to the corpse of a colleague? As in literally, because that's what the article is about: being made to work right next to the corpse

In think that's quiet extreme, honestly. Wat beyond what it'd expect any supervisor to ask of the employees.

most people have some connections to their co-workers. And if one of your friends dies right in front of you... it should be human decency to at least give then some time to settle until the body has been taken care off.


Youre responding to an account less than a month old

Ah, I didn't check that. Thanks for pointing that out.

I guess that was me interacting with a bot


Fwiw, piratebay continues to be the -to my knowledge- biggest public tracker out there, with basically every media production available.

And I think that was his point. They may ruin some people's life's... but aside from that, they achieve nothing.


Just visited two nights ago and was pleasantly surprised to see domains active with fresh magnets. The people are returning to nature.

It never stopped!

> basically every media production available.

not sure why you made this claim as its not at all germane to the discussion but this isn't even remotely true.


> piratebay continues to be the -to my knowledge- biggest public tracker out there

It has been compromised for more than a decade. The site is impossible to navigate without an adblocker due to malicious redirect ads and most of the major torrents are being monitored by rights management companies who will notify the user’s ISP of suspected infringement.


All public torrents are monitored by the movie studios, that has nothing to do with how The Pirate Bay is run. Users can just hop jurisdictions with a VPN and use these public torrents without ever getting any complaint letters.

Those are just crappy replicas of 'The Pirate Bay'. Full of annoying junk ads. 1337x is probably the biggest public tracker.

and even 1337x is too annoying to use at times compared to RuTracker

yeah, rutracker is to my knowledge the biggest and "cleanest" semi-public tracker.

I would like to mention that Google own SPA framework, angular, has redirect routes which effectively do back button hijacking if used, because they add the url you're redirecting from to the history.

https://angular.dev/guide/routing/redirecting-routes


That's only because SPAs routing maybe internal for that particular app. In that case, the UI/UX to end user is *only* possible and smooth when they hijack the back button as such.

Yeah, I think people are forgetting that if we want PWAs to behave and function like "native apps" that this functionality is needed.

I revently built an app with SvelteKit + Capacitor and had to implement the stack-based navigation one expects from a phone app myself.


getAngular was an online JSON storage service. The guys reused their code when they went to work in DoubleClick at Google. Not made by Google not Google’s framework and plenty of Google people think it’s rubbish.

While that's an interesting history tidbit, angular 2 was remade entirely under Google's leadership - the previous AngularJS framework was not reused to my knowledge - so I'd strongly push back on your statement it not being made by Google

"made by Google" was added by the 2 guys from the failed startup when Angular was first released (before Angular 2), without permission from the rest of Google.

Haven't used it like that myself, but stumbled upon this last year

https://tailscale.com/docs/concepts/domain-ownership

This let's you use your own domain for your tailnet, isn't the funnel but - but isn't it even better? Unless you actually want a publicly routable domain name, then you're back some hosted ingress I guess


And you think it's acceptable to upload photographs of your friends to some random service to use as it sees fit?

Glad I'm not your friend, honestly.


Tbf, solar has gotten so much more effective/cost efficient in the last 12-24 months that it's beating pretty much everything aside from hydro in the cost efficiency department at this point - including (most of) northern Europe and Canada.

Most data you find will be using data that's massively out of date and be off by at least 2x though...

I had another facepalm moment when I read about EU planning to go nuclear again. That would've been amazing and smart in 2015 - but now? Yeah, it's dumb af. And that's coming from a German living at the northern end of the country.


Germany spends 10x more than france on transmission and curtailment each year. Households have highest prices in EU per Eurostat despite EEG subsidies. Even if everything goes well gas expansion is still required to firm renewables. All this while it still burns coal and gas.

Going nuclear was sane in the past and sane now. If Germany wants to prove expanding nuclear is dumb it should try first to have lower annual emissions, while spending less than double the cost of entire french fleet.

France is the biggest winner in EU- it'll build both nuclear and renewables achieving deep decarbonization


EEG Subsidies no longer exist. Germany's high electricity price is due to the weird af Laws on Renewables, terrible planning, and well Gas Power, which is just expensive as shit.

eeg still exists. It just moved to state expenditure when before it was paid directly. "weird af Laws on Renewables" - which laws? Gas firming was planned long time ago and mentioned even by fraunhofer

Specifically alot of the laws on Windpower. Where it has to be amount x away from any populated place. Meaning there is only something like 0.1% of the area of Germany where you would be allowed to built Windparks.

My bad on the EEG i thought that stopped since I do not pay it anymore directly.


I doubt wind laws expansion makes it that expensive. Laws were recently liberalized anyway.

But there are many more cost factors. With co2 tax, gas and coal firming are getting very expensive. Add to that massive transmission and distribution expansion/upgrades costs for distributed grid, add massive curtailment costs and you get what you get


Can batteries store enough energy for dunkelflaute in winter? I don't think it's possible with the current technology.

Batteries are not appropriate for dealing with Dunkelflauten. There's very little energy flowing through there, so what you want to do is trade lower round trip efficiency for lower capex. The high capex of batteries is best amortized over many charge/discharge cycles, for example for daily storage.

I mean, who cares? Fire up the gas plants in the one week a year you have weather anomalies. We’d still be 90+% carbon free which would be incredible. The last gap can be solved at a later point as technology evolves

And replacing the natural gas burned in those turbines with hydrogen won't be very expensive, since they will be used so infrequently. Storing energy as hydrogen is much cheaper than storing it in batteries, as measured by cost of storage of capacity.

My friend, renewables only have a capacity factor of .1 (10%). That means those "gas plants" (really coal, and the worst quality coal on the planet too) are running 90% of the time. There is a reason why France's grid makes 7x the power for the same CO2 emissions as Germany.

A single energy source having a capacity factor of 10% does not imply that gas plants will have to run 90% of the time.

It ignores storage, over-provisioning, aggregation of uncorrelated sources etc.

Not to mention that wind typically has a much higher capacity factor than 10%.

I don't know what the true number is, but I think this is a low effort take.


Wind turbines across a whole region you'd be looking at 30% maybe 35% or even 40% if they're off-shore. Off-shore the winds aren't slowed by all the random structures humans build but also the turbines are much taller and as your elevation increases the reliability of the wind increases.

PV it varies by how far you are from the equator, 10% is realistic for a Northern country like the UK or Germany whereas in Africa you might see 25% or even 30%


It's not. Germany would need an insane amount, about 3twh based on recent data and much more looking at 30y weather data

Batteries can store as much energy as you are willing to buy.

> I had another facepalm moment when I read about EU planning to go nuclear again. That would've been amazing and smart in 2015 - but now? Yeah, it's dumb af. And that's coming from a German living at the northern end of the country.

In 2015, Germany produced about 650 TWh of electricity. In 2025, it’s around 507 TWh, a drop of roughly 22–23%.

Consumption has also declined, mainly due to efficiency improvements, higher energy prices, and weaker industrial demand.

Per person, that’s about 7,900 kWh in 2015 vs ~6,000 kWh in 2025. France is at roughly 8,000 kWh per person today, so basically where Germany used to be.

This happened despite adding about 100 TWh from wind and solar combined over the same period.

Wind is still volatile and hasn’t really ramped much in recent years, while solar is growing steadily, but mostly helps in summer.

And that’s the core issue. Solar output in summer is roughly 3× higher than in winter, so just adding more solar doesn’t solve those cold, dark winter periods without massive storage or backup.

To get back to 2015 production levels of around 650 TWh, Germany would need to increase output by about 30%. With solar growing by roughly 13–14 TWh per year and wind not increasing much recently, that puts you close to a decade just to get back to where you were, while 2030 demand is already projected at 700–750 TWh.

Given that Germany still imports around 70% of its total energy, it’s hard to call it a “facepalm” to suggest nuclear as part of the mix.

Also worth noting that Germany is still slow on smart meter rollout, with only around 2% of metering points using smart metering systems so far. That limits how much consumers can respond to real-time prices. During tight periods, this can increase reliance on imports and contribute to higher prices in connected markets such as the Nordics.


Wind has a problem in Germany thats true, but the problem is not volatility its the maddening regulations that basically only exist because nimbys do not want Windparks built anywhere.

What I'll watch which great interest is how big an improvement in interseason storage you need for the situation to flip on it's head entirely.

If sodium-ion, or some kind of thermal, or some kind of gravitationnal (except pumped hydro), or whatever techno comes up that makes it possible to handle this dunkleflaute thing (i learned that word today, love it already :) [1]), then Germany will already have the panels and windmills.

If for some reason, there is a great chemistry already advanced in the labs, is it possible that Germany buys a GWh battery before the first few EPR-2s come out of the ground ?

That's one hell of a bet to make. By refusing to reconsider nuclear, Germany is basically betting on some sort of breakthrough (or continued gas supply, which, well, is betting on geopolitics...)

So maybe "carving up mountains" isn't such a crazy plan, after all...

[1] https://en.wikipedia.org/wiki/Dunkelflaute


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: