Ahmed Mohaisen’s Post

A reminder for us all — #AI has been around for ages and sometimes an issue can be easily resolved using good old classic models!  #funny #trends #ml 

A day or so ago I was minding my own #business, scrolling through #socialmedia, to see that almost everyone was posting about a #lawsuit filed by #X (formerly #twitter) against companies that avoid #advertising on the platform. It was bizarre to read that and my first reaction was like “You can’t do that! Every company has the freedom to advertise wherever they feel like!” But then I was like okay! Let’s see if there’s any more to that? It was very hard to dig deeper as everyone on social media was #trolling and joking about the matter. Luckily I ended up finding some replies here and there that explain what’s happening. Now take this with a grain of salt as I have no clue how the #legal system works. Here goes my basic understanding: It’s true that every individual company has a saying and the #freedom on where to #advertise — no dust on that. However when #companies team up to pull their #money from a specific #platform, then they’re not acting out of business interest anymore. Rather, they’re acting with the intent to cause #economic harm for non-business reasons. This act is by itself, in many cases, #illegal. What makes it more messy, is that companies often don’t do advertising themselves but hire #agencies that do the #job for them. These agencies act as #gatekeepers and hence they can easily be blamed to be abusing their power. I’m not really sure if these #laws are only applicable to business, or if they scale to cover #politics as well. I doubt it, because sanctioning a country is often taken without hesitation and I don’t think anyone can sue against that? People also pointed out that even if the case won, then X might get some settlement from the advertisers, and there’s no guarantee that it’ll result in forcing agencies to advertise there — so it’s kinda short term winning. Others say that X should have waited for the US elections to unfold, because if Trump wins, then many companies and agencies will switch their marketing strategy to focus on whatever is trending, which may make X a hot destination for advertisers. Regardless, it’s kinda interesting to know that there’s all of these laws and I’m curious to see how this case is going to unfold!

I genuinely hope that whatever is happening with #Boeing acts as a wake-up call for an #industry that is obsessed with #speed of #delivery and #profitability at any #cost. If you don’t know the details, then rest assured, nothing really new nor surprising to mention — just a company trying to deliver faster and maximizing #profit at any cost, including setting #quality aside. Maybe what’s worth mentioning is the problems Boeing is facing today haven't originated overnight — rather it’s the smell of things that were cooking behind the scenes for a few years. Many articles on the internet do a much better job explaining the #history, but the long story short, Boeing thought, and maybe still thinking, that it’s a good idea to spend less time thinking long term as well as trying to test and verify the equipment manufactured. And those #engineers who dared to oppose the direction, they’ve been told that the door is wide open. Those who believed in the #principles they live up to decided to take on a new endeavour. Kinda split on the latter, it's a bad but also good thing to happen! I mean these big #corporations have absorbed a lotta of #talented people, the smartest people on Earth are working there. And the fact that some of them are leaving and starting their own thing is everything but a bad thing. A great example is the #BlackForestLabs — a group of experts who left #StabilityAI and worked in stealth to come a few days ago and surprise the world with a text-to-image model on par with the likes of #Midjourney. The model is called #Flux and it’s #opensource, free for everyone to use.

In just two days, two powerful language models (#llms) on par with the likes of #GPT has been #open-sourced!! 😱 Hats off to #Meta and #Mistral for releasing #llama3.1 and #large2, respectively 👏

My new favourite #quote: "The most common error of a smart engineer is to optimize a thing that should not exist" Not sure who said it, though it hits hard!

Just to clarify, the issue that put the world to its knees is not caused by #Microsoft, but from third-party #CrowdStrike software that’s widely used by many #businesses worldwide for managing the security of #Windows PCs and servers. Yet, it's alarming to see how fragile the world is! 🙈

Always wondered if it’s a matter of time until #GenZ starts to abandon its values! I mean it’s easy to aim higher when money is not an issue … #Chart showing #consumer interest in #environmental, #social, and #governance (#ESG) issues is declining across generations. by #mckinsey

As I #read and #listen to all the conversations around what’s good and bad, the #future and the past, what’s ruined and what’s intact — one #pattern keeps repeating. What we’re used to is the best and should last forever, and what’s new should never exist and must go away. The #washingtonPost did a #study, where thousands have been asked about the best #music, the best #radio show, the best #fashion .. etc. Turns out music, #movies, #entertainment, and #society in general peaked during the exact time period when you, dear reader, was a #teenager. I’ll link to the #article below, unfortunately it’s hiding behind a #paywall :( https://lnkd.in/dMaTfNVN

It’s sad to see #StabilityAI (SAI for short) going through a roller coaster of events that led many to call it the downfall of an era — and the fact that SAI doesn’t respond officially to a lot of the concerns makes it even worse, a #PR nightmare in a way! #Emad, the ex-#CEO, was the echoing name among #media outlets covering the #news — claiming to be the reason behind everything wrong with SAI, just get him out and all will be fixed. Not surprising, that wasn’t true — after his resignation, the company kept going downhill and now the fingers pointing out to greed. SAI was built by a few individuals, Emad with two German researchers who all believed in #openSource and committed to #democratizing #AI, which they did. #StableDiffusion (#SD for short) is the result of this — a text-to-image model that competes with the likes of #MidJourney and #DallE. After its overwhelming success, #investors started chiming in, #Intel with millions of dollars and big ifs — where SAI was kinda forced to use this money to buy Intel hardware to train SD models. According to #Reddit, this is where the gap started — the time that marked SAI’s focus shift, from democratizing AI to generating #revenue and being #profitable. Nothing wrong with that, but when you’re leaning too hard, things may go unexpectedly wrong. Just like many users, I asked Reddit about the alternatives — assuming that SAI is no more, what options the #community left with? And it seems that the shift will be towards models like #PixArt, #HunyuanDiT, and #LuminaNext — some are made in the West while others are in China. I genuinely hope to see SAI get back on track, learn from the mistakes and apply them ASAP, as well as to remedy its relationship with the community who helped shape the models and popularize them. I believe the fall of SAI is a lose-lose situation, for both the company and the community – the only winners are the closed-source (#proprietary) players. My hopes extend as well for a more understanding community, to understand that training these models from scratch is limited for a reason – because they need a lot of time, effort, money, and resources to do that. The people who are working on making such models have bills to pay and families to feed. 

There was this #video that went #viral of a lady who shared with her boyfriend a #clip which she thought the top voted #comments were interesting, to be surprised that her boyfriend was seeing completely different top voted comments, even though he was sitting next to her. The only difference was that he was using his own #Instagram account to read the comments. I will link to that video below for more context, though the idea of this post is to shed a bit light on something interesting, which seems to be alarming for some people. The comments seen by these two individuals seem to enforce their #biases, their own ways of viewing the world. The #algorithm gleaned into multiple factors, like their gender, likes, maybe comments — to create a version of a world you’d like to see. This version won’t be authentic without the validity of some other people, I mean we’re #social animals after all — so how about validating that by showing you the responses that align with how you’d like to view the world? It’s really interesting and concerning to see that, as for me personally, I’d use the comment section to better understand the #context so I can form my own opinion. I do that often on #Reddit and I genuinely hope that Reddit doesn’t do what Instagram seems to be doing. If such an algorithm continues to thrive, it’s just a matter of time until our own version of reality will be unchallengeable. Funny enough, a day or so before that video, I was watching an #interview about #AI #alignment — on how to make sure that the AI aligns with human’s “#values”. One of the answers was how such a task is going to be kinda impossible to follow to the T, because we, humans, are not aligned. What’s interesting is a proposed #solution where a #future with advanced #virtual reality may solve this issue — not with AI, but even better, with humanity. A future where everyone straps their #headset to create a world that aligns with everything they think is correct, fair, and absolute. Little did I know that this concept has already been put into practice by social media! ------- link to the mentioned video https://lnkd.in/dC9UUq3p