Opus 5.5 Max is my favorite, but some of the Sol/Astra 6.2 ones are pretty good.
Edit: some folks reported crashes. It was probably Gemini 4 Carbon @ High, I've put it behind a warning. If you see another one let me know (OS, browser, which bowl, etc)
Which browser / version / os? And was it any particular bowl that crashed it?
It works fine for me on Chrome, Safari and Firefox. However, some models have produced crazy resource intensive ramens before. I audit them now before pushing to make sure they're not too resource intensive
It is not clear to me how or if /teamwork is any different than Claude dynamic workflows.
Antigravity as a harness feels very behind. There's not even any auto mode. And they announced /goal long after I ever felt the need to use it in CC or Codex.
There is (--mode auto), but you can't access it without restarting the harness. Even then, you'll get plenty of mother-may-I showstoppers. I had to resort to using the --dangerously-skip-permissions option to get anything done. The good news is that Carbon is a pretty awesome model once you do get it working.
Seems likely that the people who maintain Antigravity don't use Antigravity to do it. It has too many rough edges that would be trivial to fix if someone actually noticed them.
I think it's a natural reaction to the "our AI systems are really dangerous and might kill you all and will definitely take your jobs" messages that have been coming out of the AI labs for the past few years.
They've dialed them back a bit now, but way too late. That message has sunk in.
Campaigning to prevent data centers from being built is the most actionable thing a US citizen can do, because it's a local issue (and hence something they can influence), plus they can use arguments about noise and power prices and water pollution to help make the case.
(I think most of those arguments don't stand up to scrutiny - as light industry goes data centers are pretty minor, would you prefer a concrete plant in your back yard? But I don't think the arguments are the point for a lot of people. It's about having a way to oppose AI generally.)
> think it's a natural reaction to the "our AI systems are really dangerous and might kill you all and will definitely take your jobs" messages that have been coming out of the AI labs for the past few years.
It's a natural reaction to sound, water, soil an earth pollution.
Look, I use Claude every day. It's great for my work, and it genuinely makes me more productive, and I like it.
That said, I live out in the boonies (as far east from the SF bay area as you can go before there is just nothing but the forests of the Sierras). They are talking about building a data center in Ione, a small town out here in the middle of nowhere (not quite as remote as where I am). Now, I cannot fathom why that would be a good place. There's no water, there's no infrastructure, we're not near any internet backbone. Perhaps land is cheap out here, but electricity is served by PG&E, so I don't imagine it's particularly cheap. I, and everyone I know living out here, are vehemently against building data centers out here. It just makes no sense to put them out here, especially with the wildfire risks. There's really no propaganda needed to make people hate them.
I don't see why I shouldn't be against them. It doesn't benefit me in any way, it just increases my electricity costs. Nor am I sure there is even a market for the majority of proposed data center builds, if this exponential growth/need for computing power doesn't hold we will be flushing money down the toilet, and largely on the public's dime since most of it is funded by loans and debt. Even if LLMs achieve all that is promised and more, that just means more people out of work sooner and the already powerful and wealthy gain more power and wealth than anybody else.
In some alternate universe, the AI companies could have used some of their billions to subsidize consumer accounts, or something, rather than just let 'supply and demand' raise the rates for all the people that they needed to not vote against them. why should anyone not directly benefitting from AI be charged more for their monthly keeping the lights on bill? The short-sightedness of the powerful never ceases to amaze me.
> In some alternate universe, the AI companies could have used some of their billions to subsidize consumer accounts, or something,
The "or something" should have been building out power generation and feeding back a fraction of that into the grid. At least then the electricity costs can drop for the locals.
If they proposed a data centre that has it's own power generation, and fed 20% of the electricity back into the grid, there'd be less opposition.
I have no doubt there's foreign bot armies amplifying these things, but foreign bot armies understand they're most effective when the amplify real sentiments.
And the anti-data center stuff is a very, very real sentiment.
It's both. It's not either or. It's absolutely both.
There are good reasons for people to be pissed off about data centers in the U.S.
China is absolutely funneling money to push anti-datacenter sentiment, and many of the people involved/receiving that money are not aware of where that money came from.
We are at the crux of a national security spiral in AI development that is naturally unpopular at home.
You know it's entirely possible for China to be pushing this while dropping their own fortune into research, and for everyone to lose their investments.
Don't get me wrong, it's no that the tech is useless, especially for militaries and surveillance states, but it's not the economic miracle or 'new industrial revolution' so many have gotten away with promoting. I think many here will be shocked at just how quickly a fad can die when the money implodes.
Who are the people receiving money from China to be opposed to the wasteful data center buildout, and how can I get some? I'm still waiting for my check from Russia / Iran / Qatar for being opposed to Israel's genocide campaign.
>what is the purity spiral and why is it left wing?
People trying to 1-up each other leading to the beliefs of a group being more polarized[1]. The naming doesn't imply it's a partisan phenomenon, any more than "right wing politician" implies politicians are a right wing phenomenon
There is probably a lot of overlapping factors, for example I would think that the general dislike of genAI in the art/creative scene (which skews left) would play a role.
Purely anecdotal from my point of view. The folks I see being very anti-AI are also the more left-wing folks in my bubble, who I know have gone down some of the purity spirals we've had before.
I don't know many right-wing folks. Most of the ones I do know are dealing with different issues and don't really give a toss about AI one way or the other.
However, I was living in the UK when the miners lost their livelihoods. Entire communities left to fend for themselves on diminishing benefits with no support.
I see some hypocrisy in our (knowledge worker) industry. The same sort of folks who were supporting Thatcherism and cheerfully calling themselves Yuppies, completely rejecting any attempt to unionise or join any collective bargaining efforts, are the same sort of folks (though not necessarily the same individuals) who are now complaining that they'll be left without a livelihood by AI.
Best time to unionise was 30 years ago. Second best time is now. But this is still somewhat of a radical position for a HN reader to hold.
I sometimes compare the user sentiments here and what I see and hear from my peers in the real world, and boy some people in tech are detached from day to day realities of life for the majority. Parent’s a good example.
The software industry is definitely the forerunner of the changes we'll see. It's being transformed completely, way more so than other professions. So to some extent it's not surprising that we (HN) have a different viewpoint than folks in other industries.
Yeah it's not like all these ai-tech oligarchs enabled trump to the tune of hundreds of millions of dollars each and he proceeded to disenfranchise voters, kidnap minorities, and make a mockery of environmental protections. "left wing purity spiral stuff" people are watching community members without criminal records languish in work camps.
i work in AI and i think we should hit the brakes: get a better oversight mechanism in place, help local municipalities get more in exchange for building them (no more tax incentives, NDAs, high emissions power sources), those are good places to start.
rushing ahead with the 'it might destroy us but we've gotta beat China!' is silly bullshit.
And the debt financing needs to stop.
There's a sensible path forward that doesn't involve a full stop, nor maintains the previous trajectory.
My electricity bill is 50% higher than this same month 1 year ago. My friends and neighbors in my city have said that they also have greatly increased electricity bills. Whether or not that increase is due to a specific data center, it's extremely easy to blame it on the increased electricity usage of data centers, no foreign propaganda required.
My electricity cost per kWh has just about doubled over the last 3 years, but afaik, it's mostly related to legislated changes to the energy mix and increases in fuel costs.
Fuel costs are up for lots of reasons, more demand for electricity could certainly be one of them... Inflation is certainly another component.
If the problem is demand outstripping supply then the obvious thing to insist on is that if they want to build, they should add more supply than they're consuming, so that they're fronting the money to make your electric bills go down.
Whereas instead we keep seeing proposals to ban them outright, which is the thing you'd expect to be pushed by someone with a different interest than domestic electric bills.
Right, so why is the proposal to prohibit "data centers" instead of the thing you're objecting to? Make them add more in non-fossil fuel generation than they're consuming.
We’re going to pretend that these are the same concerns as an HOA squabble about wrong roofs? I like AI as a useful tool worth investing in but let’s not be dismissive about what people are actually mad about, because this dismissive attitude is exactly why SV finds itself where it is.
How many people have actually experienced a super loud data center, or a data center that's polluting their water supply, or polluting their air? To me the opposition seems to be 99% from people who are just talking about this because it's part of the zeitgeist right now.
Even the story you linked about the water supply, it was wastewater not drinking water. I'm not saying that it's good or even okay for wastewater to have bacteria in it. But if people en masse are against data centers because they "could pollute the entire town’s water supply," based on a singular case of one data center's wastewater having some bacteria, that gives me pause. This seems more like a trendy issue than a practical one for most people.
> I'm convinced the anti data center movement is foreign propaganda
Exactly! No rational person could oppose supergeniuses like Elon Musk, Sam Altman, Marc Andreessen, etc., so if people do, the logical conclusion is that they've been manipulated by a psyop.
We still don’t know exactly what happened, but I publish software to public registries, and supply chain attacks scare the hell out of me. They also frustrate me, because most of the defenses aren't automatic and often really hard to get right. We're really failing as an industry to do better here.
This is why I started to move more and more into my tix.im environment. You can't trust anything anymore - I assume everything is compromised if not proven otherwise.
I can't tell if they're withholding access to manufacture some sense of scarcity but the delays are really prolonged. I'm nervous to use Gemini models anyway, as its attached to my Google account and I've seen too many people having their whole account banned for dubious reasons (with little recourse).
Also the whole "we're only releasing it to cyber defenders first" framing is weird, is any Gemini model an actual cyber risk?
I only use the Google models via OpenRouter. Partially as setting up directly is a pain, and keeping the Gemini CLI working was another pain but also as a buffer to protect my Google account as you mention
Could it be TPU/GPU availability? ie, is it possible that it's more profitable for their compute to be serving Anthropic models, so Gemini customers are stuck with lower compute available.
They've always been realllly slow to roll out stuff outside the org when it comes to AI. Most googlers themselves likely won't have access to this model internally.
Small sample size, but I've heard from multiple physicians that AI note taking allows them to see more patients or spend more time with them and they're very happy with those kinds of capabilities that take over toil
IMHO defaults for uv, npm, etc should all have a cooldown for 24 - 48 hours by default, you should have to go out of your way to install something newer.
We have 7d cooldown on packages on an internal pypi proxy but that blocks security updates also, which we now have to manually exclude from the cooldown.
I’m unsure how to automate this properly, maybe an LLM is the best starting point.
They definitely wouldn't be if attackers knew that it would then bypass cooldowns!
Ultimately, you need someone to verify releases and that probably means paying for it. I know Sonatype paid solutions utilise their own internal research to allow/block new releases.
No, but there's a bigger tradeoff for not upgrading. In one case it's that you're potentially exposed to a known vulnerability, vs the other case where you might just be missing out on some new or slightly reworked feature.
We need something external that actually checks to see if source code matches built artifacts, and also possibly something Llm based that checks to see if you’re making checked in code vulnerable or directly adding something malicious.
Yeah those exist… the trouble being we need something that doesn’t exist
There needs to be a service that does reproducible builds of opensource projects as they sit and publishes results for people to scrape freely.
To be robust itself against attacks, likely 3-4 services.
But then that needs funding to be sustainable … and companies are incredibly cheap. Especially right now. As such… shrug :(
Maybe something as a skill for LLMs to run that actually checks build reproduction from source for a project and all of its dependencies? But then that’s gonna burn some tokens ….
I do this for personal projects … but at work they’ve started rationing tokens… as such… idk?
I didn't realize they've all such added that, this was a change we made to our own projects last year when these first started - glad to see it's common.
Risk is ensuring enough security researchers are finding these and quarantining these before that default cooldown lapses.
Yea, I think all the tools have this now but its not a default. Install scripts should also be turned off by default and require some hassle to turn them on for the first time in a package.
The last time I looked for information, I remember seeing that the median supply chain attack was in place for multiple months to years before being discovered. Cooldowns shorter than that seem like snake oil.
No, because there are malware-finding mechanisms other than "let some unlucky guy get affected first", like automated scanners. If everyone implements a cooldown, everyone will in fact be safer.
Plus some random value slow curve distribution. Basically incremental rollout from the client side. But if it's not caught at the commit time, attacker will just wait for a certain point on the curve.
I wonder if in some cases it's not so much that people are updating too quickly, but rather, their deployment pipelines do not have pinned / frozen versions attached to their builds.
It is a money grab from Amazon I'm sure. They want to make a deal with meta to get paid for access. Maybe also worried about liability I imagine on accidental unauthorized purchases?
Returns cost money and if an agent generates more returns and hitting the margins or seller retention then that's something that's likely going to get increased scrutiny.
The generally quoted return rate is 5% to 15% (and it varies by category). If Muse identified purchases are generating returns of 10% to 30%, then amazon is probably going to crack down on it. (Amazon Sellers pay from 30% to 50% for customer returns from their own pockets. https://medium.com/@myrealprofit/how-much-do-refunds-really-... )
Opus 5.5 Max is my favorite, but some of the Sol/Astra 6.2 ones are pretty good.
Edit: some folks reported crashes. It was probably Gemini 4 Carbon @ High, I've put it behind a warning. If you see another one let me know (OS, browser, which bowl, etc)
reply