GPTs Are GPTs: An Early Look at the Labor Market Impact Potential of LLMs

I asked Chat GPT which antacid medications are contraindicated for some medication I&#x27;m on. Easily found through NICE.<p>It made up a severe risk of death taking a very common medicine combo. It was super convincing, even giving information on how long to avoid taking them together. It was pure bullshit.<p>I think as much as hyping the benefits we need to hype the flaws and dangers.<p>If the public at large learn to trust these LLMs too quickly and deeply, that&#x27;s a hard hole to dig out of. Skepticism in all information sources is a key critical thinking life skill.<p>That skill is harder to apply the more natural the information source seems and the more ambient the information.

&gt; Our findings indicate that the importance of science and critical thinking skills are strongly negatively associated with exposure, suggesting that occupations requiring these skills are less likely to be impacted by current language models. Conversely, programming and writing skills show a strong positive association with exposure, implying that occupations involving these skills are more susceptible to being influenced by language models...<p>Am I reading this correctly that the assumption here is that programming and writing skills aren&#x27;t reliant on critical thinking?<p>There is also a table which indicates exposure to LLMs in various models and it shows Mathematicians to have 100% exposure. This bit is more puzzling to me. Maybe I am misunderstanding something here.<p>edit: styling

Given that the primary author is OpenAI affiliated and some of the assumptions are far fetched (e.g., on programming - not even coding) this reads like a sponsored post pamphlet to me.

Total trash cloaked in a complicated story.<p>What they actually did is ask 5 random people to rate what thought a language model could do to help different professions. These 5 random people don&#x27;t know anything about the professions they&#x27;re rating, just what anyone off the street knows, and they know as much about GPT as anyone who has briefly played with it.<p>The title should have been &quot;We asked 5 friends to see what they thought about GPT and labor market&quot;

I think more parallels should be drawn with what we were doing before: Googling it.<p>Perhaps it&#x27;s because ChatGPT seemed to happen much more suddenly than Google became a programming resource, but we&#x27;re using them in much the same way. Asking for pre-made solutions, explanations, troubleshooting tips etc.<p>ChatGPT just does the job way better. But no-one was worried Google would put knowledge workers out of a job.

This is an incredibly interesting study and brings to light some crucial points that we, as a society, need to address sooner rather than later. As GPTs become more powerful and ubiquitous, they have the potential to reshape the labor market in ways we haven&#x27;t seen since the industrial revolution.

So, researches paid by OpenAI are going to publish a paper to convince everyone else that an OpenAI product could have some economic impact.<p>Am I the only one who sees a conflict of interest here?<p>Is this preprint already submitted for peer review or published somewhere? Or is it just an advertisement formatted in LaTeX?

We are currently surviving with a massive shortage of skilled labor, where people in the richest countries have to wait days or weeks to see a doctor, and months to see a specialist and have important medical procedures.<p>The same is true for other skilled industries, where many people are excluded from access to good resources due to their scarcity.<p>We are a long way off from having too much skill. Let’s first get to parity with humanity’s needs.

One major risk with this kind of AIs is that for the last few decades the mental image of what such AIs will be like (both by scientist and science fiction) has a _extreme serve_ mismatch with how they turned out.<p>In the past humanity people thought of them as this non-emotional purely logic driven beings and in turn problems people imagined would be such as this logic not considering emotions and emotional well being of humans or blindly (but using logic) pursing a specific goal no matter the consequences or not valuing freedom or gaining emotion. But in all that it still follows logic.<p>But now we have AIs which could have all the problems above _but doesn&#x27;t use logical thinking_ to _archive goals_. Instead it uses complex overlapping _statistical models_ to _tell a believable story_ where believable is defined by the training data which is _widely inconsistent, wrong, misleading, discriminating, emotionally charged, etc._ because it&#x27;s just scrapped from the internet. So there is _no systematic finding of goals, subgoals, plan etc_, there is _no logic_, the concept of &quot;truth&quot; simply _doesn&#x27;t exist_ for such systems etc.<p>At the same time this turned out to be &quot;often times&quot; good enough to be usable for many task and can be convincing enough to make people believe that it&#x27;s sentient.<p>But this also means it will retell common false information, misconceptions, discrimination, hatred etc. from the internet.<p>Similar it will do what people call &quot;hallucination&quot; and &quot;lying&quot; but it _not_ either of that and calling it that is misleading. Because it just doing _exactly_ what it was created for: Telling a &quot;believable&quot; story given the training data.<p>And gaslighting, misleading people and lying are a extremely deep ingrained part of the internet, i.e. a deep ingrained part of the data on which it bases what is &quot;believable&quot;.<p>And while we can add tones of bandaid on top to try to hide&#x2F;filter out such &quot;bad&quot; responses IMHO without fundamentally either changing the training data or the approach this is bound to fail while even stronger upholding a misleading illusion.

OpenAI is among the authors, so this paper can be biased.

In short: GPTs could do all the boring work for us and give us more time for things we like to do.<p>An example: Of course GPT can at some point make better music than me, but I am playing an instrument not because I want to sell the recordings, but because it is a lot of fun for me.<p>Let GPT do my taxes, then I have more time for playing. (can it do taxes finally?)

With their methodology and caveats, mathematicians have 100% exposure.

This is why my anxiety is at levels unseen before. I am a programmer who’s on a working visa. Assuming it does take my job, I have no idea what to do next… other than panic.

Given an emphasis on internalized documentation, these could be significantly better, especially over time than a lot of automated support in place for a lot of companies. As long as there is some kind of escape hatch for escalation.<p>That said, the reality is closer to the &quot;Johnny Cab&quot; from the first &quot;Total Recall&quot; movie.

With little kids I am increasingly conscious about how to best equip them for the world when they are adults. I think previous playbook will need to be redrafted. &#x27;Resilience&#x27; and &#x27;Critical Thinking&#x27; are two things I am thinking are key any others?

It is important to start thinking along the lines of the government taxing corporations using LLM to increase their profit. This additional tax revenue should be used by the Government to give a universal basic income to people losing their livelihood to LLMs.

Does somebody know where to find the raw data of exposures on all O*NET occupations?

Okay so was the original acronym for generative pre-trained transformer intentionally chosen this way as a play on words?

Anybody who actually read the paper and can share some insights in it instead of spewing opinions and anecdotes?

Critical Thinking is the way

==Occupations with no labeled exposed tasks==<p>Agricultural Equipment Operators<p>Athletes and Sports Competitors<p>Automotive Glass Installers and Repairers<p>Bus and Truck Mechanics and Diesel Engine Specialists<p>Cement Masons and Concrete Finishers<p>Cooks, Short Order<p>Cutters and Trimmers, Hand<p>Derrick Operators, Oil and Gas<p>Dining Room and Cafeteria Attendants and Bartender Helpers<p>Dishwashers<p>Dredge Operators<p>Electrical Power-Line Installers and Repairers<p>Excavating and Loading Machine and Dragline Operators,<p>Surface Mining<p>Floor Layers, Except Carpet, Wood, and Hard Tiles<p>Foundry Mold and Coremakers<p>Helpers–Brickmasons, Blockmasons, Stonemasons, and Tile and<p>Marble Setters<p>Helpers–Carpenters<p>Helpers–Painters, Paperhangers, Plasterers, and Stucco Masons<p>Helpers–Pipelayers, Plumbers, Pipefitters, and Steamfitters<p>Helpers–Roofers<p>Meat, Poultry, and Fish Cutters and Trimmers<p>Motorcycle Mechanics<p>Paving, Surfacing, and Tamping Equipment Operators<p>Pile Driver Operators<p>Pourers and Casters, Metal<p>Rail-Track Laying and Maintenance Equipment Operators<p>Refractory Materials Repairers, Except Brickmasons<p>Roof Bolters, Mining<p>Roustabouts, Oil and Gas<p>Slaughterers and Meat Packers<p>Stonemasons<p>Tapers<p>Tire Repairers and Changers<p>Wellhead Pumpers

Honestly, I feel worse for checkout girls locally - people are more often replaced by improvements to low hanging fruit, and not the emergence of apparently incomprehensively profound.<p>The truth is that we&#x27;re impressed by this because it&#x27;s magic and we don&#x27;t know how magic - but it still getting way more magic really fucking fast.<p>Right now it&#x27;s still and idiot if you&#x27;re not telling it what to do. Hypnotic babble.<p>(If you want to be scared or relieved feed it some time series stock data (if it&#x27;s good it&#x27;ll end up fighting itself, otherwise it&#x27;s either stupid or a liar).)<p>There are some uses easy use cases for it - I really wouldn&#x27;t mind a GPT bot replacing outsourced call centres.<p>Improvement of quality and an improvement in service, and no direct impact to the local economy.<p>Other than that, this all has the booming feel of every other damned tech buzz, can we just cool down on sensationalism and not try and imagine another multi-billion dollar sub-economy into existence?<p>We already did that with bitcoin, and now we&#x27;re seeing the end of the magic trick; let&#x27;s not do it again.<p>How about a calm objective and public analysis by under-excited experts before we start mining land that simply isn&#x27;t there?