I've dabbled with Qwen 3.x and Gemma 4 models a bit. They are alright but not that impressive. And my mac gets super hot if I use them for extended periods of time. It's just not very nice to use locally.

I know plenty of programmers with degrees other than computer science. Geologists, biochemists, theoretical physicists, etc. Most hard sciences involve some degree of programming at this point (usually Python). And with AI, system thinking is becoming much more relevant than deep algorithmic knowledge or math skills. Nice if you can do that stuff manually but not that essential anymore.

Also, compare this to ICE engines which experience continuous explosions, lots of mechanical parts, extreme temperature swings, etc. and still manage pretty decent durability. There's simply no base for assuming that parts like this wearing out and needing to be replaced is going to be a common thing.

CATL already put sodium ion in cheap cars. And there are other benefits to this type of battery like a wider range of operating temperatures that cover essentially all of the extreme temperatures you'd find in the arctic and the hottest deserts.

I would not be surprised to find some of these batteries in big semis a few years down the line when the cost benefits make the space/weight sacrifices worth the trade off.

But you are right that domestic and grid storage are also going to be huge use cases.

LLMs can be much better at recommendations. Honestly, Amazon needs to spank their recommendation teams into doing something productive with this. They clearly have the ability to run LLMs at scale. But their in house recommendation teams seem to be stuck in the pre LLM era and there hasn't been any material change in their very broken and underwhelming recommendations in well over a decade.

I actually dumped the list of books I've bought on Amazon over the last 15 years as a text file at some point and dumped that in ChatGPT. Quite interesting to see it pick up on my tastes. What works really well is taking a few books that you enjoy and asking it to find similar books. You need to set a few guard rails. Recommend new authors, don't recommend stuff I already have read, etc. But that's not a huge amount of context. Amazon seems incapable of doing this. It always funnels me to the same tired list of recommendations of shit I've declined to buy from them for years.

This was true long before AI. With AI the difference is just a lot bigger. It exposes team inefficiencies quite mercilessly. We have a big glaring issue with the current AI tools not being to suitable for usage by multiple users. All interactions are one on one. Which means hand offs between tools and people are bottle necked on people communicating with each other. So, any issues there with people delaying, gate keeping, etc. become very visible.

The sentiment of pushing back on AI is understandable but probably not a productive reflex. We need to find more effective ways on staying on top of massive amounts of changes. It's not going to slow down and insisting on manually reviewing all code is not going to be a long term sustainable way of developing software. It simply does not scale. I'd question the added value of manual PR reviews at this point. Are they finding real issues? Are we valuing those issues correctly? Could we come up with automated ways to find and fix those same issues? There are a lot of open questions about how we are going to do this. But no question about the notion that we need to up our game on this front.

It seems very competent at coding tasks as well. I don't think Anthropic has a huge edge on that front. It's more of a neck and neck race with proponents in both camps. I ignore most benchmarks at this point; I don't think they have much relevance for normal users.

I think it's actually necessary for both to try out different approaches. Nothing is set in stone yet when it comes to the UX of these things.

The value of Github is not just hosting projects and issue tracking but the notion that it's probably the largest professional network of developers that can interact with each other, create pull requests and issues on each other's projects, etc. There are many GH alternatives but very few with the same global network of essentially almost every developer out there.

Instead of everybody creating new accounts on each other's not quite Github clones with only a handful of users, why not improve on the state of the art here and allow people to use their user account from a server of their choice and create a proper alternative to Github?

A good analogy might be networking companies and infrastructure companies during the dot com bubble. It devalued a lot of companies but the internet stayed. A lot of dot com companies didn't make it. Much of the infrastructure investment did not go to waste, however. Nor did a the technology go away.

I think it will be the same with data centers, related infrastructure, GPU hardware, algorithms, OSS components, etc. for AI companies. More companies need that stuff than is currently available. The ones that don't make it will have a lot of assets that they can pass on to the one that still have a chance. I don't think a lot of that stuff will get decommissioned or will be underutilized. It might get a little hair cut in value though. And like during the dot com bubble, some companies actually survived and did quite well. Especially those in the business of selling shovels during a gold rush.

After the inevitable consolidation that follows the next logical stages in the hype cycle, I don't think AI will go away. It might be a bit of a bloodbath for some silicon valley investors that placed the wrong bets in the last few years. But that's the price of doing business over there. That doesn't mean it's all bad. And the smarter ones probably spread their risk enough that they still might come out looking alright.

And like with the dot com bubble, many financial types have no clue what is happening and are running around like headless chickens. Which is why they ended up sinking a lot of money in exactly the wrong things. You'd hope they would have learned something.

But articles like this suggest that that might be too much to hope. They still don't really get how technology tends to not stagnate and might continue to deliver potential for performance and cost optimization. The current level of investment is only unsustainable if that doesn't happen and nothing else changes. I don't think those kind of closed world assumptions are a safe bet at all.

That varies per region but this is a sector that is now about 2x the level of investment for fossil fuel related infrastructure. A lot of which is increasingly looking like the ROI might never happen.

The rest is just economics. We might suck at making batteries. But we suck even more at making new gas/coal plants or fueling those cost effectively.

Anyway, this stuff is being deployed by the hundreds of gwh per year now. Much of it in China but some non trivial amounts in places like Texas and California as well. As a result, the grid is actually getting more stable, not less stable.

No, but technology has moved on quite a bit. Heavy fly wheels are no longer the state of the art here.

A few tens of GW of battery capacity (i.e. a few dozen nuclear plants worth of capacity) that can switch on/off in milliseconds can do a lot for grid stability. That's part of the reason why grid operators are rolling out so much batteries. It's not necessarily about supplying energy for a very long time but about smoothing out peaks and dips in energy supply and demand and responding more or less in real time to that.

This stuff is basically being rolled out at industrial scale in a lot of places. Australia, China, etc. pretty much run increasingly on mostly renewables. This is no longer as speculative as it would have been ten years ago.

Yes, there are engineering challenges with rolling that stuff out in a lot of places. And even more policy and regulation challenges. Actually that is, by far, the #1 challenge in places like the US and Europe. Grid operators are simply structured and incentivized wrong to deal with this stuff efficiently. Texas is actually not doing too bad relative to e.g. California. But they clearly have some challenges still.

The nice thing with data centers is that they are somewhat flexible. It's not a constant load. Data center operators can choose to reduce load. And if properly engineered, they could do so automatically based on signals from the grid.

The issue with outdated grids is that it relies on technology (spinning mass) that's at this point a century old. Which makes it brittle against outages like you describe. The solution is not more spinning mass but batteries and renewables to take the place of that spinning mass. A battery can respond to oscillations in milliseconds. If you then add flexible load that can spin up/down based on the amount of available power, you gain a lot of stability.

We're talking about vocal minorities expressing their arguments for and against AI. And some people here are just very vocal and dominant. Passive aggressive/obnoxious even. But that doesn't mean they represent the dominant opinion. If you've had the pleasure of attending a lot of meetings with developers, many of them barely open their mouths and some of them can't shut up. You find a lot of those types on HN. And looking in the mirror, that's me.

Most people that come here are hackers. Many of those probably use paid subscriptions for agentic coding tools at this point. At least, I don't know many professional developers that don't. But I know plenty that grumble a lot about AI and how they are still relevant and not that easily replaced by a tool. This stuff is a bit threatening to many people and it's triggering anxiety, uncertainty, anger, etc. And a lot of that leaks through in the discussions here.

And there's an uncomfortable reality that HN has been around for a while and the demographics are maybe a bit skewed towards people in their forties and fifties. People that are a bit set in their ways and maybe a bit conservative and change resistant. I'm in my fifties myself and I have to actively fight that tendency in myself.

So, all of that adds up to a lot of negativity.

The good news in a desert: plenty of sunshine. So you can generate a lot of electricity with some cheap solar panels, there is plenty of space to put some down, and there aren't a lot of NIMBYs around to complicate the permitting process for that.

Some desert ecosystems actually depend on condensation with specialized plants and animals harvesting humidity from ocean breeze. Large parts of e.g. the Sahara border on the Atlantic ocean. Lots of water in the air but not a lot of rain. And even if humidity is low, there still is some water in the air usually.

But the simple fact of course is that there is a lot more water in water than there is in air. If you want to extract meaningful amounts of water from air, you need to process a lot of it.

There's a growing cost of living and poverty crisis in some countries that probably is strongly correlated with education levels. That's also the urgent issue to address.

And there are issues with students not finishing school. Or students entering higher education without basic skills for math and literacy after actually completing high school. I know some Dutch universities have had to skill up students on basic high school math, for example. No longer being taught adequately, apparently.

And then separately you might wonder which skills are actually still relevant for people to have. People not speaking more than one language used to be a big problem in some countries. These days that's still not great but something you can compensate with using AI translations. Being able to calculate numbers is nice. But it's not the end of the world if people use a calculator for some things. But it is an issue if that's not a thing they can do.

Education was never about enlightenment and more about making sure workers were ready for a productive live factories and offices and making sure companies had access to people with a good base level education. Before the industrial revolution, most people would not spend a lot of time, or any amount of time, in schools.

But most of the cloud providers now offer Valkey because of the license changes. Of course, cloud providers not offering Redis was the intention of the license change from the Redis point of view. So mission accomplished for Redis.

But the flip side of course is that if you want to deploy on standard infrastructure rather than self hosting Redis, Valkey is now the easy, low risk path that probably should be the default for most companies that target AWS, Azure, GCP, etc. Same with Elasticsearch vs. Opensearch and a few other products where the community forked because of license changes.

Mentioning Elasticsearch because I know people in both communities and I'm deeply familiar with the stack. A few years on, Opensearch has taken a lot of the momentum from Elasticsearch.

If you are ever on a project where somebody goes "Somebody should write some documentation for X", you should counter it with "Great idea, get on it!". Mostly nothing will happen. It's rather thankless work. Some people are more proactive on this.

I actually tend to write documentation for myself. Because I'm old and wise enough to realize that if I come back to a project in two years, I will have forgotten most that I would need to get back up to speed quickly.

With agentic coding tools, it's different. The documentation helps. And it gets added to even if you don't ask for it. Which is nice. And you can get a lot of documentation added with a few simple prompts. Which makes it cheap and easy to generate.