Tech predictions for 2024 and beyond

As long as AI is controlled by Silicon Valley, it will never become culturally aware, because the people controlling it are not the slightest culturally aware. They think diversity sits in the skin colour, and that&#x27;s about as far as they are willing to think about it, because they are intellectually lazy and have dead souls, without any curiosity for the world.<p>They train their AI on some false assumptions that fits their political faith, and foreigners who need to use their AIs will have to learn about the biases and work around them. Just like it is today.

I used to enjoy <i>I, Cringely&#x27;s</i> yearly predictions.<p>He&#x27;s had medical problems and has gone silent the last 6 months or so.<p><a href="https:&#x2F;&#x2F;www.cringely.com&#x2F;category&#x2F;predictions&#x2F;" rel="nofollow noreferrer">https:&#x2F;&#x2F;www.cringely.com&#x2F;category&#x2F;predictions&#x2F;</a>

This is interesting honorific, many tech sector workers have doctor degrees, but nobody ever calls them doctors.<p>But curiously it is used in the title, which doesn&#x27;t follow HN title guidelines (i.e. editiorized from original article title).

I struggled to get through the culturally aware section. Apparently the technology should define its responses in terms of what ever identity group the user has been lumped into. For some reason if it doesn&#x27;t do this, it won&#x27;t have the same reach.

It’s one year only, I know we can’t expect a breakthrough. But this is very boring.

I remember some visions of the future from Microsoft from the early 2000s. They even demoed some cool gear with the Kinect and I even managed to try out the table at their campus.<p><a href="https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=6VfpVYYQzHs" rel="nofollow noreferrer">https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=6VfpVYYQzHs</a><p>That and this vision too.<p><a href="https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=wraF2DjALls" rel="nofollow noreferrer">https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=wraF2DjALls</a><p>Now I speak to others SWEs in tech from semis to hyperscalers and it feels so bleak. All ethereal bollocks with crypto and AI. All designed to make someone else money. Where is the cool wearable, IoT, actual human interaction tech.<p>We should be having magic mirrors, that style of coffee table... Heck I would settle on smarter green tech.<p>Everything honestly seemed more fun back then from the state of the web, gaming.<p>I have never been so bored with the state of tech.

Here are some additional tech predictions relevant to the HN community to compare and contrast:<p>Gartner tech trends: <a href="https:&#x2F;&#x2F;www.gartner.com&#x2F;en&#x2F;articles&#x2F;gartner-top-10-strategic-technology-trends-for-2024" rel="nofollow noreferrer">https:&#x2F;&#x2F;www.gartner.com&#x2F;en&#x2F;articles&#x2F;gartner-top-10-strategic...</a><p>Gartner AI predictions: <a href="https:&#x2F;&#x2F;www.cio.inc&#x2F;future-ai-gartners-predictions-for-2024-a-23601" rel="nofollow noreferrer">https:&#x2F;&#x2F;www.cio.inc&#x2F;future-ai-gartners-predictions-for-2024-...</a><p>Dell predictions: <a href="https:&#x2F;&#x2F;www.dell.com&#x2F;en-us&#x2F;perspectives&#x2F;top-tech-predictions-for-2024&#x2F;" rel="nofollow noreferrer">https:&#x2F;&#x2F;www.dell.com&#x2F;en-us&#x2F;perspectives&#x2F;top-tech-predictions...</a><p>Deloitte predictions: <a href="https:&#x2F;&#x2F;www.deloitte.com&#x2F;global&#x2F;en&#x2F;about&#x2F;press-room&#x2F;deloitte-technology-media-and-telecom-2024-predictions.html" rel="nofollow noreferrer">https:&#x2F;&#x2F;www.deloitte.com&#x2F;global&#x2F;en&#x2F;about&#x2F;press-room&#x2F;deloitte...</a><p>BDO predictions: <a href="https:&#x2F;&#x2F;www.bdo.com&#x2F;insights&#x2F;industries&#x2F;technology&#x2F;bdo-tech-predictions-for-2024" rel="nofollow noreferrer">https:&#x2F;&#x2F;www.bdo.com&#x2F;insights&#x2F;industries&#x2F;technology&#x2F;bdo-tech-...</a><p>NTT predictions: <a href="https:&#x2F;&#x2F;finance.yahoo.com&#x2F;news&#x2F;ntt-announces-key-technology-predictions-150000358.html" rel="nofollow noreferrer">https:&#x2F;&#x2F;finance.yahoo.com&#x2F;news&#x2F;ntt-announces-key-technology-...</a><p>Techopedia: <a href="https:&#x2F;&#x2F;www.techopedia.com&#x2F;tech-layoffs-predictions" rel="nofollow noreferrer">https:&#x2F;&#x2F;www.techopedia.com&#x2F;tech-layoffs-predictions</a><p>MIT Tech Review: <a href="https:&#x2F;&#x2F;www.technologyreview.com&#x2F;2022&#x2F;02&#x2F;23&#x2F;1045416&#x2F;10-breakthrough-technologies-2022&#x2F;" rel="nofollow noreferrer">https:&#x2F;&#x2F;www.technologyreview.com&#x2F;2022&#x2F;02&#x2F;23&#x2F;1045416&#x2F;10-break...</a>

&gt; Culture influences everything<p>While culture definitely touches everything and is a lens through which we look at the world and each other, there are some things that culture cannot bend to its will. I&#x27;m not arguing that there is an objective reality that we can access, but there are things in the universe that culture cannot make a dent in.<p>I had a conversation just this morning with Bard where it provided me with a culture slant that is also, shall we say, politically influenced, but in this particular instance I just wanted the raw facts without taking people feelings into account. It felt like it was lecturing me, which is very very off-putting.

Is even a single one of these concrete enough to be verifiable? Is this what passes for hypotheses in our business?<p>If you make predictions, <i>please</i> make them concrete enough (actual numbers!) that we can tell in hindsight whether you were right or wrong. Without being able to judge someone&#x27;s prediction track record, it&#x27;s all useless for informing any real decision.

I&#x27;m expecting a sobering year for AI in 2024, a result fueled by marketing gimmicks, closed training data and weights, false promises, chaos in recruiting, more tech layoffs due to economy (and false hopes for AI-powered workforce replacement), demotivation in the young workforce, etc.<p>Another reality hit for AI implementation will be the cost of generative AI API services and environmental impact of the process of training the models.<p>But I am not a Doctor, and CTO of big monopolist and I don&#x27;t have a stake to sell you AI cloud services. So who knows..

Here&#x27;s my prediction: 2025 predictions by the Amazon CTO will not be written by the Amazon CTO, but by an LLM.

&gt; Unburdened by the undifferentiated heavy lifting of tasks like upgrading Java versions, developers can focus on the creative work that drives innovation.<p>I totally agree that AI is going to make it easier and easier for engineers to get closer to the customer and work on real problems. The number of specialist &quot;IT&quot; people, or separate backend and frontend classifications, will decrease and the startup and product world will greatly benefit.

If you didn&#x27;t have LLMs on your list of tech predictions a year or two ago why should you have any credibility in predicting the next breakthrough?

Tech predictions? More like tech wishful thinking

Ok literally none of his predictions have any merit. They are absolutely bullshit. My bullshit detector has not gone higher this month. No wonder he is the CTO of Amazon. All of these companies are on their downhill dump.

Yeah, there is always the high-level view of these things and the low-level.<p>The way LLMs make stuff up is oft-remarked upon and is unclear whether it&#x27;s a dealbreaker or not but it might be for any&#x2F;all of this to come to pass.<p>Here&#x27;s an Werner&#x2F;AWS-specific LLM anecdote along those lines:<p>I went onto AWS&#x27;s LLM Q for the first time today, a little chat icon nicely&#x2F;helpfully linked from their documentation pages (when logged into my AWS account) and asked it a question which I was unable to find an answer in their docs or via Google&#x2F;Bing&#x2F;StackOverflow after 5-10 minutes of searching.<p><pre><code> What Redshift SQL query can I write to tell whether the instance I am connected to is a Redshift Serverless or a Redshift Provisioned cluster? </code></pre> (I have migrated Redshift clusters in Dev&#x2F;Nonprod to Serverless to reduce cost, but featureset is not identical so some of our app code is sensitive to the differences and needs to detect which type of environment is present before executing. The release number, but not the instance type, can be found in Redshift SELECT VERSION();.)<p>AWS Q LLM replies:<p><pre><code> To check if an Amazon Redshift cluster is provisioned or serverless, you can run the following SQL query: SELECT cluster_type FROM svv_cluster_info; The cluster_type will return either &quot;provisioned&quot; or &quot;serverless&quot;. You can only connect to and query Serverless clusters using the Amazon Redshift Query Editor v2. Provisioned clusters can be queried using either Query Editor v1 or v2. [...+more text on the differences and a bulleted list of 3 weblinks&#x2F;sources provided...] </code></pre> Wow! Awesome! This answered my question perfectly! The response and code is formatted beautifully just like real documentation. There are a lot of SVV_ system views and I guess I just missed this one. I am at this point very impressed.<p>I go to actually try the above query on my serverless and provisioned clusters...<p><pre><code> Serverless: SQL Error [42P01]: ERROR: relation &quot;svv_cluster_info&quot; does not exist Provisioned: SQL Error [42P01]: ERROR: relation &quot;svv_cluster_info&quot; does not exist </code></pre> Aww, man!<p>Another attempt answered similarly affirmatively, this time mentioning an imaginary &quot;SVV_CLUSTER&quot; view (which also didn&#x27;t work.) So I don&#x27;t think this is just some &quot;unreleased future feature&quot; leaked to the LLM via its training set of internal documentation (wouldn&#x27;t that be nice! err, kinda...), since even the SVV_ view name isn&#x27;t the same when you ask a similar question at different times.<p>Even as a chatbot, the most obvious use case, this is is not really meeting the mark.<p>I know it&#x27;s brand new and rushed out of the door for re:Invent two weeks ago and I know it&#x27;ll get better... but it was a bit hard to read Werner&#x27;s article a few hours later and take it seriously. My first Bayesian prior is now a &quot;err, no&quot;.

lol his 2023 predictions were way off..just saying

this all sounds great for rich people and like hell for everyone else. AI learns that brown people exist, women become data farms, clippy is your new boss, children are our future profits.<p>There is some kind of satire in here about how tech leaders and QAnon folks want the same things but take opposing routes to get there. Tech wants to say the N-word, get inside women’s private parts, devalue knowledge work, and indoctrinate children. But for liberalism…