stefangordon

I tend to find this part of the work enjoyable. It is just a faster and more productive version of what I would do with any engineer working for me who owns a large feature. I barely have to hint at my concern or drop the right keyword and the agent (Fable/Astra) will immediately understand.

"Couldn't this be stateless?" "Do you have a plan to be able to shard this?" - We will almost always pivot from the agents initial design but the agent is able to easily understand the reasons and benefits and align quickly.

If I was writing the code myself, I'd often have to make compromises between the ideal architecture and the level of effort required to implement it - now I can just always have the ideal architecture.

NASA, with the best leadership it has ever had, is cutting absurdly wasteful, over-budget programs that will never launch, to focus on actually getting things done.

Sounds great to me.

The devices are entirely open for all practical purposes - but worrying about individual cameras is silly, because they have no meaningful security at all around the API's to access all the cloud data - you can buy law enforcement credentials dirt cheap in dark web marketplaces to log in and track anyone/anywhere you want and access all footage.

Our agents automatically review our PR's - the authors agents automatically see the feedback and make fixes, and automatically merge when everything is green.

A well authored CI review process is significantly better than any human could do. We have the AI review not only the changes but clone and investigate all related repositories that integrate with the code to evaluate interactions, check all open and historical issues that may be related, review production/development logs for the feature, etc.

It will then also go update issues/tasks to reflect the changes after the merge.

Similarly you can have automated agents watching post-deployment to mitigate risk in the "a new feature just went out" time-frame. Just having someone check the related logs in the hours after deployment can catch a lot of issues that otherwise would turn into a fire-drill.

I can't tell if this is an excellent joke, or a sad not-joke.

“constant missing in our math” is roughly how inventing new particles in physics works.

When your equations are missing a number to work, you announce a new particle.

Free Opus 5 is like getting a free STD.

It dramatically improved about a week ago - most of my blocked projects are now completely functional.

It wasn't really worth exploring last year either.

For your use case n8n may be a good fit, and Claude or Codex will be happy to set it up for you.

If you want this "messenger interface with computers for the agents" feature set Grok Bot is quite good and effortless to use. With Tailscale configured on its personal VM it can connect to machines on your network easily.

The exponential growth they are having right now is absolutely insane. I would caution against assuming its an easy problem, or that anyone else has demonstrated an ability to handle this level of scale/growth any better.

In May they had to solve for 30x growth overnight, and I would guess its closer to 100x by now. The reality is a centralized solution like this may no longer be a viable option with the scale of code generation.

It is only unreasonable if you have the unreasonable assumption that this is not a simulation.

In my experience the complexity of the coding challenges tackled correlates well with the talent of the engineers.

It is uncommon to find engineers doing work they find trivial outside of some consulting niches.

Indeed. I find it quite bizarre that just about the only thing the LLM appears incapable of doing is writing appropriate instructions for an LLM - whether it be memories, sub agent prompts, planning documents.

I’m struggling to imagine a project that would require more than one talented engineer and a bucket of tokens anymore.

I think perhaps the assumption that engineering managers should have any employees may be outdated.

I can imagine average and mediocre engineers equipped with tokens could create chaos and debt on a scale never before imaginable, so it’s easy to see how orgs who still have these employees around are struggling with the transition.

The reality is you need to get rid of them all, and replace them with the most experienced highest paid person you can find. In the near future that person will become obsolete too.

They are concerned mathematicians are obsolete and seem to be unaware that super intelligence means humans are obsolete.

You need to be thinking about an entirely different form of life for all of humanity, happening in the very near future. Focusing on one specific field that may be earlier in the obsolescence chain is a distraction.

Merit based society based on intelligence or work of nearly any kind is about to cease existence.

For some reason it feels to me that this should make mathematicians feels better. :)

I would learn both Claude and ChatGPT just to be safe.

Clearly you didn't try it yet ;)

I'm an Opus fanboy, but this is literally the worst coding model I have used in 6 months. Its completely unusable and borderline dangerous. It appears to think less than haiku, will take any sort of absurd shortcut to achieve its goal, refuses to do any reasoning. I was back on 4.6 within 2 hours.

Did Anthropic just give up their entire momentum on this garbage in an effort to increase profitability?

People worried about routers, meanwhile nearly every damn employee at Intel from the CEO to the janitor is Chinese.

The Intel ME chip is running its own OS on every single Intel chipset, even when the computer or laptop is shut down, and accessible directly through attached Intel WiFi or network cards. With full memory access, with no way to turn it off.

https://en.wikipedia.org/wiki/Intel_Management_Engine

The totality of reassurance we have about it is intel’s promise that they won’t put a backdoor in.

That is what TPLink PR would like you to think.

The reality is the only part that matters, the chipsets, are produced in Chinese factories owned by TPLink.

They moved everything that doesn’t matter to the US recently in an effort to give the illusion that they aren’t putting chips manufactured under the control of the Chinese government into the majority of routers used in the US.

I’m not agreeing with banning them, but I can certainly see how it creates significant risks that I would want to mitigate somehow.

The relevant link is probably "Evidence for diverse structural correlation networks from the ABCD study" https://www.sciencedirect.com/science/article/pii/S105381191...

But has anyone found the full text of this yet?