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

Given that Mailchimp has now received the Bending Spoons kiss of death, what are easy-to-migrate-to alternatives?

Mailchimp has been kind of a dumpster fire for years. Major competitor-wise, Constant Contact?

Slightly worried about geohot given this and other doom-laden posts. He feels a little terminally online.

yes and no. i am worried about anyone right now that spends alot of time researching these doom topics. its a bit like climate change research. if you get deep into it it all seems terrible.

In the end we do not know all the variables and in a chaotic system even a tiny measument error will lead to very different outcomes. so we better not only doom.

Expect the worst, hope for the best.


> a bit like climate change research. if you get deep into it it all seems terrible

No it's not, climate change research is about real, present, and indeed terrible things; AI doomerism is all hypotheses.


put it another way: I used to trick or treat in light snow. now I'm wearing a t-shirt to rake the last of my leaves.

AI has yet to really replace anyone on my team and the long term usability of the AI code is dubious.


Because seasons have been shifted by about 30 days..

Look at heat maps from 15 or 20 years ago. You can see pretty much the same temperatures as now.

Not saying climate change doesn’t exist, but this is not how you measure it


give us sources..

What? Has the end of winter shifted?

yes, you heard it. by 30 days... earlier

Thank you. AI doomerism drives me nuts for lots of reasons, but chief among them is how it seems to displace people’s concern for real, existing problems with concern for completely hypothetical ones.

We know the Earth is warming at an unprecedented pace and that greenhouse gas emissions are accelerating that trend. Coral is bleaching, kelp forests are collapsing, polar bears are drowning, forests are burning...we're on a trajectory for camels in Canada; whether that is in 50 years or 100 years is hard to say, but the trend is undeniable.

RSI is just a hypothetical. It's not something that we know is imminently possible, let alone something happening right now.


AI is also a climate catastrophe, many are powered by methane peaker plants and have energy demands commensurate with medium sized cities.

Thank you. Climate doomerism drives me nuts for lots of reasons, but chief among them is how it seems to displace people’s concern for real, existing problems with concern for projected future ones.

We know that energy poverty harms people right now and that affordable energy is essential to improving their lives. Families can’t afford heating, hospitals need reliable electricity, farmers need fertilizer, workers need transportation...for many people, fossil fuels underpin the path out of poverty. Whether affordable alternatives can replace them in 20 years or 50 years is hard to say, but the need exists today.

The most catastrophic climate scenarios are projections. It doesn't make the economic plight of people who depend on fossil fuels something we can afford to ignore.

/s Congrats, you're the same as climate denialism of 20 years ago. You're the same people, unable to see the future issues heading towards us when they're not imminently clear.


Renewables are cheaper than fossil fuels today, and the path out of poverty is renewables. Remove fossil fuel subsidies and take away tarrifs on Chinese solar, the energy market will sort itself out.

Yes, if you change all of the facts, AI doom and climate change are the same.

Climate change is based on fundamental properties of physics and chemistry that we have understood for more than a hundred years. Our understanding of how the planet will change as it heats up is still limited, but our understanding of climate change itself is and has always been based on real, observed phenomena.

RSI is a thought experiment.


?

That bs about the climate was in jest?


At this point I almost forgot what he was actually good at, because it's certainly not monthly superficial drama posts. Where is the cool tech stuff he used to do?

I suppose that FOSS self-driving-car framework stuff is genuinely cool tech.

I mean, they did add a GPU slot to a car. That alone is very cool.

Not sure how much he is still doing the heavy lifting there tho.


> Not sure how much he is still doing the heavy lifting there tho.

Not sure how much ~he~ anyone is still doing the heavy lifting there tho.


A little bit of ignorance keeps the doctor away.

Always has been

He does seem that way. 5 months earlier, he was dooming on AI but from the Ed Zitron side. He explicitly made a prediction that AI would never get good enough for coding.

https://geohot.github.io/blog/jekyll/update/2026/05/24/the-e...

He now believes in the x risk doomsday lesswrongish predictions.

5 months?! You can do a complete reversal? Bro went from Ed Zitron to Nate Soares in 5 months. WTF.

I think he just likes culture wars


A lot of people changed their mind on AI, not because it's good now, but because "it makes stuff up, does shoddy work, and wastes time" apparently isn't a showstopper. It is being pushed everywhere even though it never got good. If you were thinking "this stuff sucks, people will get bored soon and move on" and then you saw everyone using it despite it sucking, then you have to reconsider the consequences. You can lose your job even if AI doesn't do it as well. Your free software codebases can all be tainted. It didn't have to get good in the end. Only popular.

This is not relevant here because the author believes RSI is imminent and used for world domination. Just a few months after he said that coding will never be solved by agents

I don't think anyone has spelled out what RSI [actually] is this whole thread, and I didn't see it in the article either. Apologies if I've ignored a key element here.

I'm not sure if you are asking sincerely or not but the Author has mentioned it

> After RSI comes, all the demand will be internally generated, agents working with other agents, everyone else cut out of the loop except them.

RSI = recursive self improvement. Without going into details, you can't hold this view (that RSI is imminent) and that agents can't do coding.


I read the article very differently from you. My takeaway is that geohot thinks RSI is bullshit but enough powerful people believe it is imminent that the mere conjecture of RSI will be used as a cassus belli for a protracted world war with China and friends.

Could have changed his mind? I'm sure he's using AI coding tools and maybe he sees how advanced they have become..

This is what he wrote, judge it for yourself

    Without fully endorsing all their ideas, I’m now in the LeCun/Marcus camp on LLMs. I don’t think models like this will ever be able to program, I think the process matters. I think that deep learning is still the solution, but real programming agents will need world models, not some RLVR shit that comments out the failing test and tells you all the tests are now passing.

    The real story of this era will be who manages to avoid harming themselves in their AI psychosis.

I’m sure he sees the irony in the last sentence

There may be growing evidence now, such as OpenAI's math releases on September 8 and October 6. Soon, these advances could be applied to R&D and help solve hard technical problems.

It seems to take a while for quite a lot of people to properly go into a topic.

I'm following AI and ML for years by now which means reading something about it nearly every day, following progress on different topics like LLM but also other architectures, research results, ML projects etc.

The ignorant representation of LLM is just a parrot has been going through HN for ages and you might still find someone today saying the same thing.

If you don't do any experimentation with LLM and just use it for coding on one code base you might get the feeling that its good and thats it. If you do experiments and you see an LLM control Blender on your cmputer and BUILDINg a real complex 3D thingy, it feels complelty different.

If you have your LLM and 20$ subscription on a powerless notebook it feels compmlelty different than a 100$ subscription on a machine with GPUs. Why? Suddenly you don't limit yourself due to subscription limit and with GPUs now your Claude can do physics simulations and optimization problems and now you gave it a lot better of tools to use.

The progress is staggering and while others complain that HN is full of annoying AI news, they just miss complelty whats going on in th eindustry.

This is the most craziest time ever. I have never seen any technology to continue progressing like this.


Or, more generously, dude overcorrected after it became clear that his prior beliefs were utterly wrong. Doesn't seem like a particularly nuts chain of logic, when presented with irrefutable evidence that AI can surpass your expectations, to assume it can go far beyond them.

We'd be clowning on anybody who did the opposite, constantly correcting their predictions by just a little goal post moving: "Computers will never pass the Turing Test, wait, no, computers will never be able to engage in even basic problem solving, wait no, computers will never be able to write computer programs, wait no..." etc.


I like how you’ve put it but I’d say that moving goalposts slightly each time is normal. But changing your views 180 in 5 months is not a good sign.

The right thing this guy should do is to sit back and contemplate for a bit before posting confidentially incorrect things.

But sure, if benefit of doubt is to be given then yeah he’s over correcting again


September was a big deal

Claudeslopped design and writing style. Pass.

So disconnected from reality as to be difficult to ascertain whether this is satirical fiction.

I seem to remember a few years back this guy made a lot of noise about a "free energy" (ahem) device he'd invented and claimed he was giving the plans to China and Russia as a way of getting back at the UK government for some slight or other, then promptly dropped off the face of the internet in protest(?) I guess he's back.

IIRC it worked by using a big gyroscope to stay stationary while the earth rotated, causing 1 rotation every 24 hours. So not technically free energy, just taking it from the earth

I was quite worried that he'd gone off the deep end when he published that video


The way I understood his "free energy"-device back in the day, was supposedly trying to induce a current into a set of coils from the very tiny discrepancies between the earth's rotation and that of its magnetic field. How the gyroscope would play into it, is anybody's guess (alignment, I suppose?). Obviously, none of this would work, and I don't think he was in a particularly good place back then.

I saw an artwork like this before, maybe by Anish Kapur?

I remember him showing a cube with an outlet and a switch that could power a lightbulb. He eventually revealed that inside were just a bunch of lithium ion cells taken out of disposable vapes

That’s not a very charitable interpretation. He was distraught about not being able to get his wife into the UK and was lashing out. He removed the video, he knows it was silly. Hopefully the situation was sorted.

It's just what I remembered, clearly I had no recollection of that part. My sympathies with him if he was in some kind of strife.

His wife is mentioned as being behind the camera in the latest video, so it sounds like it was sorted.

Yeah UK marriage visas are fucking expensive, long and humiliating.

Its one of those things that was borne out of tabloid panic (IMPORTING WIVES AND FAMILIES!!) which means that now everyone has to go through pain until a scandal or another panic causes a change.


It's exactly the same as the green card problem in the US. The UK is not exclusively bad here.

I don't know the specifics of the UK marriage visa process but as someone who went through the American one, I feel like it's not that bad in terms of visas. There's the usual bureaucracy but nothing terrible. Obviously my experience is Trump 1 through Biden administrations, Trump 2 has made it worse.

Whatever came of his hydrogen/electrolysis project? He has talked before about how those big control panel boxes he made and the high voltage transformers were to contol some kind of high power electrolysis rig that would make a huge hot-air-balloon-like flame. Not sure if that was some kind of free energy effort or if he just wanted to make fire with electricity.

A stoichiometric mix of hydrogen and oxygen, which is what you get when you electrolyse water, will straight up detonate when ignited. A normal party balloon of the gas is like a gunshot when it pops. A 1meter diameter balloon will probably wake up a whole city.

He probably did it at a small scale, and realized how phenomenally stupid it is and declined to pursue the idea further.

Further reading: https://youtu.be/l9CI6KSV560?is=R8LTPX9esprEqmCp


Claudeslopped site design and information layout makes me dubious of the information conveyed.


The animal cartoons are also incredibly off-putting. I don't get it.


I close as soon as I see them, or a back and forth with questions between them.


So many exclamation points!


Zoom/scroll/pan is utterly broken. Which is a shame as the idea is novel, if only it weren't Claudeslopped.


I'm not sure I'd call it grim, per se. Blood Meridian takes that label for me, almost at the exclusion of all else in what could loosely be called the "western" genre. Lonesome Dove is more just very... gritty. And packs a serious emotional heft. Great book either way.


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

Search: