It all depends on how you prompt. and the prompt system you’ve setup.. when done well, you just “steer” the code /system. Quite amazing to see it come together. But there are multiple layers to this.
Yes, I personally think so. In the hands of an experienced user you can crank out work that would take days or weeks even, and get to the meat of the problem you care about much quicker. Just churning out bespoke boilerplate code is a massive time saver, as is using LLMs to narrow in on docs, features etc. Even high level mathematicians are beginning to incorporate LLM use (early days though).
I cant think of an example where an LLM will get in the way of 90% of the stuff people do. The 10% will always be bespoke and need a human to drive forward as they are the ones that create demand for the code / work etc.
The problem is many users are not experienced. And the more they rely on AI to do their work, the less likely they are to ever become experienced.
An inexperienced junior engineer delegating all their work to an LLM is an absolute recipe for disaster, both for the coworkers and product. Code reviews take at least 3x as long. They cannot justify their decisions because the decisions aren't theirs. I've seen it first hand.
I agree totally; most people are no experienced, and there is a weird situation where the productivity gains are bifurcated. I have also seen a lot of developers unable to steer the LLM as they can’t pick up on issues they would otherwise have learned through experience. Interesting to see what will happen but probably gonna be a shit show for younger devs.
It seems you've registered this account a couple of months ago only to basically repeat this opinion over and over (sprinkled with some anti-science opinions on top).
great engineering effort was spent to make software at FAANG built on clear service oriented modular architectures, and thus easy to develop for. Add to that good organization of process where engineers spend most of their time doing actual dev work.
Enterprise software is different beast - large fragile [quasi]monoliths, good luck for [current] AI to make a meaningful fixes and/or feature development in it. And even if AI manages to speed up actual development multiple times, the impact would be still small as actual development takes relatively small share of overall work in enterprise software. Of course it will come here too, just somewhat later than at places like FAANG.
No it doesn’t. I’m not trying to make a point about vaccines, just that the study is a population study and so shows benefits on average to a population.
If the vaccine killed 1/100 people (again I don’t believe this but it’s the internet) but made the other 99 immune to dying over the 4 years, it would look really good on average even if it was directly responsible for the deaths of 1%.
This comment helps me understand how folks see "your taxes will go up $10k but you won't pay $20k in health insurance premiums" as a hit to the pocketbook.
Well, if say the vaccine gave 1/100 fatal lung cancer then a population study would show a decrease in covid deaths and an increase in lung cancer deaths though.
It's only the case if the vaccine gave everybody slightly higher chances of dying from everything that it could hide in the weeds.
So in this specific example we can see from Table 2 that deaths/1 million are just lower for everything in the vaccinated so it's not the case that it lowered one kind of death drastically at the expense of another.
Don't those 99 enjoy being alive despite all of the things that would have killed some of them had they not taken the vaccine? If "some" is at least 1%, that sounds like an individual benefit to me.
If you take the vaccine, you have a lower chance of dying over those 4 years. You also have an infinitely higher chance (specifically 1% vs 0%) of dying from the vaccine, but that doesn't change the previous sentence.
For vaccines like the measles vaccine where it can entirely stop the spread in a vaccinated population this can be true until enough people think this way that measles starts spreading in your vicinity.
But with Covid-19 vaccination wasn't able to eliminate its spread so it mostly is about protecting yourself rather than protecting others.
Whats funny is in some sense, temporal replaces alot of the AWS stack. You dont really need queues, step functions, lambdas, and the rest. I personally think its a better compute model than the wildly complicated AWS infra. Deploying temporal on compute primitives is simply better, and allows for you to be cloud agnostic.
I sometimes suspect AWS deliberately looks for ways to extract low-overhead tasks into dedicated services for the simple reason that many people will pay for the service without thinking about whether they really need it.
If you bring something down in a real way, you can forget about someone trusting you with a big project in the future. You basically need to switch orgs
AWS has been in long term decline, most of the platform is just in keeping the lights on mode. Its also why they are behind on AI, alot of would be innovative employees get crushed under red tape and performance management
There is almost no reason to delegate the work, especially low level grunt work.
People disputing this are either in denial, or lacking the skill set to leverage AI.
One or two more Opus releases from anthropic and this field is cooked
reply