It heavily depends on the task, language and experience. What I saw in my team was that low performers started being as productive as mediocre programmers, however they mostly produce slop which needs to be sent back for rework, so in the short term it's not that great. However since they can iterate faster, I have the feeling that they are learning faster so hopefully in the long run they can be somewhat productive.
The high-performers on the other hand are ripping through tasks at such a speed that they now have time to focus on quality and processes, so in the end you get a really well polished piece of software. However this is starting to change for the worse as the PMs are realizing what's happening and cramming more and more stuff in the sprints.
The amazing part is not the above though - it's that one needs not be an SME anymore. One of our customers asked us to make an Android application with some features of our product, which were not trivial to port. Even though we never did anything like that before, a small team was able to release it in a couple of months mostly bug-free. Without AI this would have taken at least half a year, or more. It is quite shocking to hear the status updates going from "we need this thing implemented and I have no idea what it even means" to "it's done and works" the next day.
The stripey wood-and-black-rubber 'sound dampening' panels are all the rage these days and I don't really understand it.
Our management had the bright idea to put these things in all our meeting rooms on the wall with the TVs we use for remote calls. People started getting sea sick looking at them. Of course removing them would mean the management made a mistake, so they will stay there until the next bright idea hits.
Right up until one tries to set up a self-hosted server (spoiler - you can't, at least not without 'significant effort' - they themselves say that if you ask about it).
Eaton and APC at least have models with LFP chemistry, with comparable prices across power ratings. The LFP will be more expensive though due to the increased longevity, at least until lead-acid ones stops being produced.
The biggest issue with these diy builds is that they need the mechanics to be inside the lens - which is not necessarily a bad thing, however it severely limits the lens choice for the system, and introduces additional cost (you basically need to buy a shutter with every lens). The scene has definitely improved over the years, there are a number of very interesting x-pan-like builds which have been made possible by advancements in 3d printing.
I'm looking forward to the day someone figures out how to modify a full frame shutter assembly (plenty and cheap on ebay) to work with medium format film.
Is this really what they use to train commanding officers? It has all the hallmarks of a self-help book - vague advice coupled with some anecdotes - with a lot of bureaucratic fluff inbetween. How/why did the squad leader 'instinctively know' how to handle the reluctant machine gunner? Isn't that the opposite of training military personnel?
The other day I was trying to figure out a cost estimate for Gemini 2.5 Pro but they don't have this in their pricing calculator - it just isn't there. Only endless tables with various model sizes and options which aren't explained.
A noob question, since the original article doesn't go into details - what is exactly being simulated here? I was under the impression that we can't even reliably do a single protein folding due to the sheer complexity of the task. So how do we simulate the zillions that are bouncing around in a single cell? And if we don't simulate it at that level, how are we confident that it is correct?
Every time I try breathwork (be it box or x-y-z) I feel that the intervals are too long - by the time I finish breathing out my brain goes into panic mode and the next breath is not enough to compensate. I find them the exact opposite of relaxing.
Pandora's box is now open and multiple groups have access to drones[0].
This is something that I think escapes engineers in this line of work - that something they invented will eventually end up (legally or not) in the hands of people with no scruples.
I think many in this thread are underestimating the desire of VPs and CTOs to just offload the risk somewhere else. Quite a lot of companies handling sensitive data are already using various services in the cloud and it hasn't been a problem before - even in Europe with its GDPR laws. Just sign an NDA or whatever with OpenAI/Google/etc. and if any data gets leaked they are on the hook.
How does one even begin to do that? Looking at people I know who describe themselves as "truth-seeking", it seems that it is a one way ticket to Conspiracyland.
> Europe’s dependence on the United States for its security means that the United States possesses a de facto veto on the direction of European defense. Since the 1990s, the United States has typically used its effective veto power to block the defense ambitions of the European Union. This has frequently resulted in an absurd situation where Washington loudly insists that Europe do more on defense but then strongly objects when Europe’s political union—the European Union—tries to answer the call. This policy approach has been a grand strategic error—one that has weakened NATO militarily, strained the trans-Atlantic alliance, and contributed to the relative decline in Europe’s global clout. As a result, one of America’s closest partners and allies of first resort is not nearly as powerful as it could be.
The high-performers on the other hand are ripping through tasks at such a speed that they now have time to focus on quality and processes, so in the end you get a really well polished piece of software. However this is starting to change for the worse as the PMs are realizing what's happening and cramming more and more stuff in the sprints.
The amazing part is not the above though - it's that one needs not be an SME anymore. One of our customers asked us to make an Android application with some features of our product, which were not trivial to port. Even though we never did anything like that before, a small team was able to release it in a couple of months mostly bug-free. Without AI this would have taken at least half a year, or more. It is quite shocking to hear the status updates going from "we need this thing implemented and I have no idea what it even means" to "it's done and works" the next day.