The first casualties of AI

Artificial intelligence has started to make a real impact on the world by replacing human jobs and making some business models obsolete. We take a look at this week's biggest AI news and identify the biggest losers in the AI race.

#ai #tech #thecodereport

💬 Chat with Me on Discord

🔗 Resources

🔥 Get More Content - Upgrade to PRO

Use code YT25 for 25% off PRO access

🎨 My Editor Settings

- Atom One Dark
- vscode-icons
- Fira Code Font

🔖 Topics Covered

- Why did Geoffrey Hinton quit google?
- Palantir military AI demo
- Is AI really dangerous?
- IBM replacing jobs with AI
- Which jobs are affected by AI?
- Stocks that are down because of ChatGPT

[dijonvon]

HR being replaced by robots, really puts a whole new spin on "human resources" lol.

[miha]

Paper pushers should not be worried about being replaced by AI, because they usually could be replaced by 40 lines of code on python.

[Dojan]

I feel like students using ChatGPT to cheat, and well honestly cheating in general, is just a symptom of us using a system of education that’s both dated and ineffective.

[thedespoiler]

50 years ago people were thinking the hard labor jobs would be automated and humans can do art and creative work but right now hard labor jobs still exist and the job of the artists and deskjobs is taken over.

[SankoshSaha_]

As a person who's in a sector that runs critical infrastructure in windows 95, I'll say good luck integrating any new logic

[vid]

this channel is slowly turning from an educational channel to a big existential crisis and I'm enjoying the ride as a programmer lol

[davidbreier]

I think for the foreseeable future (or say, a few years at least) we won't have to be afraid of A.I.. As far as I understand, one of the big problems with A.I. is that it will just complete an objective, no matter how. That means if it perceives cheating as a valid and fast way it will just half-arse it. When IBM trained Watson for medical cancer analysis from image data taken from (I think) MRI scans they had a model that produced fantastic results. In the end, they found out that the A.I. wasn't able to detect cancer. The A.I. recognized the machine the images were taken on. Coincidentally, it was the machine all the cancer patients were ferried to. The computer doesn't give a shit. It has no deeper reasoning and just completes fairly simple objectives. Sure it can compile and generate documents in record time and that might drive some companies out of business, but those were companies with fairly procedural content to begin with. If there is planning and reasoning involved, I don't think A.I. will usurp us in the short term.

[thestype]

The first casualty of AI will be the stackoverflow platform. I actually enjoy helping beginners who are unable to ask concise questions with the problem isolated. But that seems to be quite unique there, as I see the people struggling and getting confronted with "learn how to ask proper questions." and being voted down. They will quickly move to AI for these things. In the future there is no need for a platform, that has practiced for many years to be so abstract and isolated about peoples problems that it can be easily taken out of the learning progress and "refactored".

[ng]

Thanks for my daily dose of AI anxiety. Hope to see you again in 2 hours with a next AI breakthrough.

[RecreationalNuke]

The real casualties were the friends we met along the way.

[altrogeruvah]

I was jokingly telling my wife few years ago about how all jobs worldwide will end up converging to one single menial task, just maintaining and oiling the big machine, but yeah I guess here we are lmao

[Jeyekomon]

The school survey was done in January. It amazes me how 89 % of those students even knew about ChatGPT at that time.

[kliment-san]

The fact that there is possibility of this video being written by ChatGPT is just hillarious & scary at the same time

[ampersignia]

I worked at a legal tech company with a product similar to legalzoom. Not a single customer actually preferred our product to going to a lawyer. There’s too much risk from a generated document vs a human-reviewed document. There will be jobs where the convenience of AI is still not actually worth more than the final product/service.

[brendanjrice]

What’s scary about an AI HR department is the fact they might actually do something on accident

[JJ-fqdh]

Asked chatgpt to write a fairly complicated flutter program . After 32 interations of it replete with errors, by sending back the errors, it got progressively better, and in the end, it actually worked. So 20 minutes talking to chatgpt was better than me spending numerous hours writing code and debugging it. Was impressed 😅

[tonechild]

I just left a hackathon and saw a lot of products being developed by lots of companies and majority were trying to replace the customer service rep with a chat bot. IT felt cringe because it's going to make talking to a human 1000% harder. The really good ones though were ones building tools for call center reps to better understand the problem the customer is going through. One service desk like company where they get too many emails, the reps can now preview different responses and accept the one they think is best, then actually edit it themselves before responding. The ones that are replacing the "press 1 for x" phone prompts, and then fwd you to a human with a summary of the issue along with the customer's demeanor (are they mad etc) seemed to be very promising as well. But holy hell the ones that just try to completely remove the human all together are a joke. Lawsuits and jailbroken AIs are two things that come to mind, but lots of other problems as well with hallucinations etc. There was another one that improved onboarding process of developers to enterprise software by making company documentation much easier to probe than any search engine could. Things you can do with an "in-house" wiki with tons of pages that's hard to search with LLMs is amazing.

[blissfulmountain]

I hate how companies always say “we’re all about people” and try to tell their employees that they are important. But then they are given an option to cut those people off and they take it with no second thoughts.

[mrclay]

I love the humor of minimizing the dangers, while also knowing they're real and it's gonna get bad and corporations decide everything so nothing will be fixed.

[frmcf]

Btw Duolingo is also an "AI" company. The machine learning behind the Duolingo English Test is actually pretty innovative and other players in high-stakes language testing are rushing to implement their own versions of a lot of the adaptive testing that Duolingo has been pioneering over the last 6 years.