If you want to be AI proof you need to learn to work with AI

As a junior developer ChatGPT has honestly increased my productivity by over 80%. I can def see a business case if they want to reduce the number of junior positions in our team. Why does it mean for the future of folks in my field I don't know?

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:People really, really overstate the power of AI. The potential is there but there’s a long way to go.

Besides, CEOs aren’t going to put them and their buds out of work.

You come across like the CEO of a buggy-whip manufacturer.

PP is right. AI has a lot of potential but has a long way to go, including overcoming major physical/practical barriers. I’m in the industry and expectations/hype in no way matches the short/medium term use cases.

Nobody in the industry thinks this.

Completely untrue. The people who are actually building products with AI are very realistic about where the tech is, where it’s going, and how long it will take to get there. The companies with the AI algorithms are the ones selling dreams.

What specifically are you trying to implement? Maybe your team just sucks?

No one on my team — a group that implements AI solutions for Federal clients — remotely feels that hype is outpacing results in AI implementations.

I recently worked on deployment of NLP semantic search for a law enforcement client, and everyone from investigators to senior IT management was blown away by the results. My firm has a portfolio of similar wildly successful deployments in everything from cleared stuff to consumer-facing implementations.

Lol sure, government contractor. My team at a fortune 100 company might suck - and You keep thinking AI is going to fundamentally remake the labor force in the near term. Blowing away an IT administrator at a federal agency is not that impressive high water mark.

Bless your heart. The fact that you talk about “AI algorithms” shows that you’re a boomer who’s never been within 100 miles of an AI project, but keep LARPing a techie of DCUM if that floats your boat.

Meanwhile, people who know anything at all about defense and intelligence know exactly who is driving AI progress.

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:People really, really overstate the power of AI. The potential is there but there’s a long way to go.

Besides, CEOs aren’t going to put them and their buds out of work.

You come across like the CEO of a buggy-whip manufacturer.

PP is right. AI has a lot of potential but has a long way to go, including overcoming major physical/practical barriers. I’m in the industry and expectations/hype in no way matches the short/medium term use cases.

Nobody in the industry thinks this.

Completely untrue. The people who are actually building products with AI are very realistic about where the tech is, where it’s going, and how long it will take to get there. The companies with the AI algorithms are the ones selling dreams.

What specifically are you trying to implement? Maybe your team just sucks?

No one on my team — a group that implements AI solutions for Federal clients — remotely feels that hype is outpacing results in AI implementations.

I recently worked on deployment of NLP semantic search for a law enforcement client, and everyone from investigators to senior IT management was blown away by the results. My firm has a portfolio of similar wildly successful deployments in everything from cleared stuff to consumer-facing implementations.

Lol sure, government contractor. My team at a fortune 100 company might suck - and You keep thinking AI is going to fundamentally remake the labor force in the near term. Blowing away an IT administrator at a federal agency is not that impressive high water mark.

Anonymous wrote:I have a doctorate too and have worked in an AI-focused company in tech. I still think we need another breakthrough before it gets really serious, but I could see it happening in the next decade or two.

I regularly dream of going to pastry school. But it's too physically demanding. I have bakers in my family and I know it's hard and very unglamorous work. Still. I prefer to be screen-free.

Baking is so precise. Robots can probably do it better than human.

Anonymous wrote:Can you train AI to pluck all the stupid dandelions on my lawn.

I recall reading an article 10 years ago, when someone invented a machine with ML vision capabilities to spray with herbicides weeds in lettuce fields. I have no doubts it will have no problems with dandelions. I am only not sure if it ever went into mass production - the accuracy was a concern for a long time.

Anonymous wrote:I find myself wanting to do something AI or a robot can't ever do, besides birthing a baby.
I have a doctorate.
I am tired of spending so much time online.
I am thinking of other options besides remote or in office positions.

I am sorry to tell you this, but it might be too late for you to become a lawn mower operator.
Your PhD is likely to be a major turn off for many owners of landscaping businesses.

Plus I am fairly sure ML can be relatively easily incorporated into Cub Cadets. They have done it in Silicon Valley with other agricultural equipment.

You are a scientist, right? So you are not afraid of an intellectual challenge! Harness AI in your field and discover new, exciting uncharted territories. Look at it as a competition - you vs. AI. Right now AI still only aspires to one day become as smart as an average IQ college grad (LLMs). You are already way ahead of it. And if you use AI as your tool to supercharge your growth - you will likely remain ahead of it for a long time.

That being said, if you truly want to make a career change, and AI is a concern, then I recommend that you look for professions that use/produce very little data (text or numbers) and compensations are moderate to low. These are likely to be the most difficult to convert to AI (AI models are very data dependant) and private sector owners/executives will have least incentive to invest resources into transition (cost savings from eliminating high paying jobs).

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:People really, really overstate the power of AI. The potential is there but there’s a long way to go.

Besides, CEOs aren’t going to put them and their buds out of work.

You come across like the CEO of a buggy-whip manufacturer.

PP is right. AI has a lot of potential but has a long way to go, including overcoming major physical/practical barriers. I’m in the industry and expectations/hype in no way matches the short/medium term use cases.

Nobody in the industry thinks this.

Completely untrue. The people who are actually building products with AI are very realistic about where the tech is, where it’s going, and how long it will take to get there. The companies with the AI algorithms are the ones selling dreams.

What specifically are you trying to implement? Maybe your team just sucks?

No one on my team — a group that implements AI solutions for Federal clients — remotely feels that hype is outpacing results in AI implementations.

I recently worked on deployment of NLP semantic search for a law enforcement client, and everyone from investigators to senior IT management was blown away by the results. My firm has a portfolio of similar wildly successful deployments in everything from cleared stuff to consumer-facing implementations.

As a fed I'm so pessimistic about this. There are so many manual process steps that I totally assumed would already be automated by regular, non-AI software (and could be if it weren't for the many human approvals, accountability checks, and layers of security from both IT and regulatory requirements). We have an entire division to manage these functions that has grown in the last 5 years as we've adopted new software that has just multiplied the steps. If Congress is going to continually increase the levels of accountability and cybersecurity that never decrease the amount of work surrounding the core function, we aren't in danger of being replaced during the rest of my career.

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:People really, really overstate the power of AI. The potential is there but there’s a long way to go.

Besides, CEOs aren’t going to put them and their buds out of work.

You come across like the CEO of a buggy-whip manufacturer.

PP is right. AI has a lot of potential but has a long way to go, including overcoming major physical/practical barriers. I’m in the industry and expectations/hype in no way matches the short/medium term use cases.

Nobody in the industry thinks this.

Completely untrue. The people who are actually building products with AI are very realistic about where the tech is, where it’s going, and how long it will take to get there. The companies with the AI algorithms are the ones selling dreams.

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:People really, really overstate the power of AI. The potential is there but there’s a long way to go.

Besides, CEOs aren’t going to put them and their buds out of work.

You come across like the CEO of a buggy-whip manufacturer.

PP is right. AI has a lot of potential but has a long way to go, including overcoming major physical/practical barriers. I’m in the industry and expectations/hype in no way matches the short/medium term use cases.

AI is currently far, far exceeding the hype. Studies continue to find that the number of “secret cyborgs” (people who do most of all of their job using AI) is skyrocketing and even so estimates likely vastly underreport.

And the reasons are obvious - as the owner of a small software product company, a combination of ChatGPT and Copilot can do more, better, with less direction for me than 5-6 junior developers that I have to hire, manage and train. I’m far more productive and follow best practices more now on my own with AI tools (e.g. complete test coverage, far more re usability, loose coupling, SOLID principals, etc.) than with employees working for me previously.

Yeah but junior developers are doing extremely basic tasks most of the time, so it's not surprising, though I know a lot of devs who say they spend more time debugging when they GPT-generated code.
The real question is... who's going to be doing the more advanced stuff in the long run?

Yes, junior developers tend to do repetitive, basic stuff. What I’m saying is that in my business I no longer need them. That’s a problem for anybody who eventually wants to be a senior dev because they need a junior developer job right out of college.

Yes, AI code assistants produce work of wildly varying quality and must always be reviewed line by line. The point is though, and even with those limitations, I alone can do the work that I used to do with a team of 5. And the end result is faster, better quality, and has 90%+ test coverage.

That’s the state of play today. In the future, fewer and fewer developers will be able to do more and more with less human help.

Remember - the AI you’re using today will be the worst AI you will ever use, by a wide margin. And that’s true every day for the foreseeable future.

I am absolutely exploring ways to make money other than spending so many hours online. Screen work is beginning to wear me down. Hello "ego depletion".

I also find myself making more and more errors. I have no interest in online meetings and I don't see myself doing desk work for 17 more years until retirement.

I am in a very strange place. I have thought about finishing my Ph.D in Counseling and finding work in mental health.

I just told my daughter she should skip college and learn how a heat pump works. Then build a business on that knowledge.

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:People really, really overstate the power of AI. The potential is there but there’s a long way to go.

Besides, CEOs aren’t going to put them and their buds out of work.

You come across like the CEO of a buggy-whip manufacturer.

PP is right. AI has a lot of potential but has a long way to go, including overcoming major physical/practical barriers. I’m in the industry and expectations/hype in no way matches the short/medium term use cases.

Nobody in the industry thinks this.

+1
People are confusing AI, AGI, and robotics.

You have maybe ten years before AI is a really serious threat. Use that time to save as much money as you can for retirement and pay down your mortgage.