TyroneLT (@TyroneLT) header
TyroneLT (@TyroneLT) profile picture

Video Creator // Video Producer I create things to help people look up 🇨🇦 + 🇹🇹 + 🇯🇲 + ➕ = ME

Ajax, Ontario
Joined January 2009
Photo by TyroneLT (@TyroneLT)Photo by TyroneLT (@TyroneLT)Photo by TyroneLT (@TyroneLT)Photo by TyroneLT (@TyroneLT)Photo by TyroneLT (@TyroneLT)Photo by TyroneLT (@TyroneLT)Photo by TyroneLT (@TyroneLT)Photo by TyroneLT (@TyroneLT)Photo by TyroneLT (@TyroneLT)Photo by TyroneLT (@TyroneLT)
Content consumed without content processed is useless.
1
1
1
6
I have a major result to share. I’ve found conclusive forensic evidence that the top 4 Chinese AI labs—DeepSeek, Moonshot, MiniMax and Ant Group—have built their newest models on NVIDIA’s most advanced Blackwell chips, which are barred from export to China. This sounds insane, but I discovered that almost every AI model carries a hidden signature of the GPU that trained it. When a model is first set up, the starting values of its weights are laid out in a pattern that depends on the exact GPU. Each combination of GPU and seed number produces a unique pattern—a sort of fingerprint of the hardware used to create the model. Traces of that pattern survive training into the published model. Again, this evidence isn’t based on any insider information or leaking. It’s based on a forensic analysis of the model weights released by the companies themselves. I searched more than six trillion combinations of chip layout and seed number and matched them to open-weights models. In doing so, I found conclusive evidence that six Chinese AI models—including DeepSeek-V4-Pro, Kimi K3, and MiniMax-M3—were initialized on NVIDIA Blackwell systems. Four of the models match Grace Blackwell, the GB200 and GB300 parts in NVIDIA’s flagship liquid-cooled rack-scale systems. I then verified these results on real hardware. Running unmodified PyTorch with the seeds I’d recovered, a Blackwell B300 reproduced DeepSeek-V4-Flash’s untouched starting weights bit for bit, and a rented GB200 reproduced Kimi K3’s to a correlation of 0.999995. A Hopper H100 reproduced none of DeepSeek-V4-Flash’s, and changing the seed by one broke the match. (DeepSeek’s seed is 953049583 by the way.) The method also was verified to match 24 known open-weights models to the hardware that developers said they used. It is exceedingly unlikely that these results are a coincidence. Across six trillion combinations, the best accidental match should reach only about 8 sigma (i.e., eight standard deviations above what random chance would produce). The real matches range from 49 to 1,385 sigma, and even the weakest single test hits 20. These were data-center-scale training runs. With some analysis of release schedules and model details, I estimate that the largest, Kimi K3 and DeepSeek-V4-Pro, used 2,000 to 3,000 Blackwell GPUs each. Clusters of this size cost on the order of $150 to $200 million each to buy, and across these labs alone would add up to more than half a billion dollars of export-controlled GPUs, not counting experimental compute or otherwise. It’s quite likely that we see multi-billion dollar expenditure on export-controlled GPUs. I am highly confident in this result, and it has been independently verified by senior researchers at OpenAI and Anthropic. You can read the full report at unemploymentlabs.com/checkpo…. There is much more I’ve identified about these models through other forms of checkpoint forensics like this, and will have more to share soon.
12
Ragebait x1000
Dario Amodei says cancer, Alzheimer’s, heart disease, and mental illness will be relics of the past in less than a decade, and that "we can create an age of miracles" “You know, if today we look at things like cancer, Alzheimer’s, heart disease, mental illness, these will be relics of the past in, I think, less than a decade." "And I believe that the company which gets the risks right will also be the company that has the license to operate that society trusts and society believes can build the benefits in the right way." "I think if we do this right, we can create an age of miracles where humanity is able to solve problems that have eluded us for millennia." "We have to make sure that all the risks, all the ways that things can go wrong with this technology, we very carefully address those risks as we bring the benefits, and as we have a viable business at the same time.”
25
This is why I stay in the following feed and then use the search function when I want to see anything else. The For You tab is a rage bait rabbit hole.
The X algorithm is getting so dumb lately. Click on one stupid bait post and your entire feed is already ruined. Watch just one video, and suddenly you get nothing but random AI video clips.
22
My feed is filled bots and paid AI bros crying because the Dems are gonna pause AI and regulate it. The AI bubble's about to pop anyway, so might as well slow it all down and regulate it before it makes a bigger mess and makes everything cost even more.
18
"Stealing like an artist" is not the same as copying with AI.
17
The sound every Big Tech CEO is likely going to hear now that the Democrats have control of congress. 😅
53
What’s bad about this? The most beneficial parts of AI can move forward regardless of legislation. W Without gigantic money burning data centres and can run on narrow AIs that focus on specific purposes and not just trying to build digital God.
This is incredibly bad news. Abdul El-Sayed wants to pause AI, while at the same time, almost every AI data center is being blocked by local communities. Folks fail to realize just how transformative AI is for humanity. It’s a shame.
23
This is what happens when AI Slop flyers mutate into AI Slop video. Average and mediocre content. Bursting at the seams, flooding our feeds more than ever before.
IMO this is the biggest shift in online video - everyone AI one-click copying. It's nearly impossible to know who made the original. I'm guessing people have agents set up to do it automatically when a concept takes off. Views are increasingly meaningless and paying creators will probably become near impossible? Can't build a business around view metrics.
19
This is what I was afraid of. Much like AI Slop flyers, we now are gonna see a similar kind of slop via low-effort AI edited video flooding our feeds. 🥴🥴🥴
im starting to notice editslop all around me 🤨
1
1
233
The most nuanced POV on this AI editing topic I have heard yet. 👏🏿👏🏿
When it comes to the AI Video Editing conversation with Claude taking over Adobe and DaVinci I think it’s actually very simple for people to just tell the truth and stop hiding behind emotional and moral arguments… That ultimately are a costume for job security or artistic preference. 80% of formats can be handled by this workflow and should be. There are plenty of genres of content there the “human” part or craft part has almost NOTHING to do with the edit are all done in camera and in front of the camera and are either personality driven or fully driven by the content itself. News videos are mechanical, not necessarily the most fun thing in the world to edit and don’t have novelty if you’re actually delivering news. They are driven by tight deadlines and non artistic edits and are more listened to than watch. The same for podcast interviews and podcasts clips. This work is basically just a slog and not something people wake up passionate r edit. If anything they are passionate to film it instead and the editing just gets in the way. I think most people would agree to that. Yes, someone could make it artsy but that’s not really the point. I think everyone can also concede that for smaller channels doing personal content that hiring an editor and also working on the edits yourself is not necessarily in the budget for MOST small channels that barely make money or can be monetized. If people can’t look at the current economy and admit that, I think we are talking past each other and have to go back to what “objectively means” when it concerns the majority. For most of those folks this will not only edit faster than them, but better and their time is better spent being good on camera, scriptwriting and putting the energy into their ideas, and any money they make back into their bills or the content itself in terms of any materials or production. The best example of this is DIY content, makeup tutorials and get ready with me content and unboxing or haul videos. Those videos are not driven by the edit or by editing based storytelling. It’s driven by personality, product and the process being shown. The marker for those videos isn’t looking to the edit as the substance. And for the creators in those niches you have to concede the budget is in the materials. Here are the genres AI Video Editing can’t touch and shouldn’t touch unless you just want the rough cut out of the way… (pauses and silences and multicam): - Pure Scripted Entertainment Content - Video Essays and Narrative Content - Cinematic Vlogs and Storytelling - Comedy and Sketches - Science Communicator Videos - Social Experiments and Pranks - Gaming Content - Storytelling and Spectacle Videos - Music Videos and Parody And here is the truth, for those of you who are video editors, that’s the content you enjoy editing anyway. You didn’t get into editing to edit B2B talking head content or Get Ready with Me Videos. You didn’t get into video editing excited to edit another political pundit interview where you have to sit there and do multi-can and audio clean up. You don’t get excited and feel the need to stamp your “craft” on a top 10 listicle video about free software… The “cool” and “fun” stuff you want to edit by hand is exactly the thing that should be edited by hand by a human. And AI doesn’t do a good job with it as things stand and in the near future of the next 5 years I don’t think it will be able to do more than help tweak with some of these genres or do the mechanical parts of them. Nobody is saying you can’t feel threatened, but the way to react to a threat to your job is to show up to your job as a professional… not crash out like a child… and concede the facts as the y are even if they point to truths you don’t particularly care for. There are niches of content where video editing is part of the value proposition. And the reality is also that MOST content creators are not your customer. Focus on the ones that are…
1
116
You're cherry picking. I guess you forgot when Star Trek did the same thing with the Enterprise crew and everyone loved it.
"Disney never gave us big hero shots of the big three in Star Wars" Yea, remember when a director got fired from Star Wars after releasing one of the worst movies ever made and then went and made a Jurassic Park sequel that did exactly that and it sucked?
45
TyroneLT retweeted
"Disney never gave us big hero shots of the big three in Star Wars" Yea, remember when a director got fired from Star Wars after releasing one of the worst movies ever made and then went and made a Jurassic Park sequel that did exactly that and it sucked?
34
9
5
157
20,991
There is risk to prioritizing minimizing risk.
34
All those Avengers Doomsday haters gonna hate all the way to the box office to buy their tickets and sit in the front row with their crocodile tears.
1
120
You don't get what you deserve. You get the results of your actions.
1
1
1
65
Being with God > Doing for God
1
51
The answer is foldables. There are already foldable phones that allow the back camera to be used while still being able to see the screen.
I wonder if Apple will ever acknowledge this and give us the second screen we all crave
1
73
This person acts like video games and junk food haven't had their day in court. It's social media's turn for better or worse. Personally I'd love for us to go back to algorhtim-less feeds by default. Let people use the search function if they want to find something.
This is disgusting and I can’t wait for the appeals. The precedent set by YouTube being liable for screen-time addiction is kind of scary. Treating algorithms like a defective product opens the door to endless lawsuits over “addictive” tech. What’s next? Books, video games, junk food? What happened to personal responsibility?
58
What this theory doesn't account for is the evolution of the internet. The net was far smaller when the prequels came out and didn't have a mature attention based business model yet. Sequel haters will grift off thst business model and polarization for the foreseeable future.
This quoted post is unavailable.
70
An obsession with distraction has become the nature of modern life.
1
1
51