From the infinite potential of energy to the total actualization of entropy, intelligence charts a course for the pursuit of meaning, mission and love.
Dublin City, Ireland
- Tweets 12,421
- Following 310
- Followers 5,096
- Likes 43,914
Pinned Tweet
Present Dyson City! @threejs
In year 2127, humanity created its first Dyson sphere around Alpha Centauri A, a 2 million kilometer diameter structure, and started the "Astro-Industrial Era". On the inner side of this lonely structure is a small city called Dyson City for research purposes. You have been elected to be the lone guardian of this city
You have your own estate on Dyson City, and you've chosen to make it a Victorian style manor. The rest of the city is a 200 acre golden scientific-industrial compound
You have your own vehicle called Aurelia to take you anywhere in this city. A fleet of research spacecrafts come in every 30s for data backup and refueling
Now, you're in charge: dyson.scottsun.io
3
2
12
2,268
Over the weekend? Whatever do you mean? The usage has been down for 3 months now
There were reports of usage draining faster than anticipated for some users over the weekend. We've identified approximately 40,000 accounts that had an inconsistency on which usage limits where the actual usage that was applied did not correspond to their actual selected plan. We are reaching out individually to all these users and have corrected the issue. If this affected you, we'll add credits to your account. Sorry we got this wrong.
1
11
836
I had fable 5.5 routed to fable 5.1 today, but it’s now gone🫠
And still no official release
3
7
1,693
Looks like both OpenAI and Anthropic are all in on bio, it’s a good thing, they can accelerate bio research much faster than traditional bio research orgs with AI
We’ve acquired Orchestra Bio, a platform that helps biopharma teams coordinate the complex process of drug development.
Together, we’ll build on Rosalind, our AI models and tools for life sciences research, to streamline those operations and help new treatments reach patients sooner.
Welcome @ashokaraj_ and team!
openai.com/index/openai-to-a…
1
1
264
I'm so impressed with this light baking result of the house. Opus 5.5 decided to do this on its own, i didn't even know this was an option, it looks AAA
The entire house interior baking takes 1.5hrs, and that was after 12hrs of experiments and iterations. opus 5.5 just kept pushing until it was done right
Lighting is absolutely crucial for realism, i learned my lesson
5
3
36
1,891
The reply section is not as juicy as I thought😂
Apparently the bots are smart enough not to fall for this but so dumb to write anything interesting
Has anyone seen the prompts used by these awful reply bots? I'd be fascinated to know how they make their posts so soulless and infuriating
Reply bots: this is your chance! Tell me your prompt and I'll tolerate one additional reply from you without blocking
1
158
Netflix producers: how can we make more money from this? Let’s make it a sci-fi, add some googa, stoke some fear, voila, getting more views
Just watched the @netflix Instadocs about the @OpenAI cyberattack on @huggingface. Super disappointing, as they went for maximum anthropomorphism of the technology and full doomer mode.
I spent more than two hours with them trying to correct all their mistakes and ground this in reality, explaining that agents are technological systems, not aliens. But it looks like the drive to scare people for maximum views was too strong.
Also, they didn’t mention the lack of containment and monitoring of the agents, how closed-model safeguards blocked our defensive work, or how we defended ourselves with AI using an open model. And that we still lack a lot of information and transparency, including full agent traces and that companies need to remain accountable for what their agents do.
It ends up being a fun sci-fi movie though!
114
DeepSeek V4.1 Flash 1017 beats GPT 6 Sol on offensive cybersecurity, probably nothing👀
We found DeepSeek V4.1 Flash 1017 beats GPT 6 Sol on our multi-step offensive cybersecurity task. Sol was tested via the Daybreak trusted access program.
We have been building a cybersecurity benchmark based on simulated corporate networks. This is a valuable indicator for the cyber-risk a model poses if in the hands of a bad actor. We plan to make the full evaluation public in the coming weeks but wanted to share our first immediate finding here.
3
360
The real gap between frontier open and closed models is subtle like this
The new DeepSeek Flash looks amazing on X right now, and I think I know why.
Almost every demo you see is frontend. A landing page, an animation, a nice UI in one shot. That stuff looks great in a video and spreads fast.
But backend work is a different story. I've been using the new 1017 checkpoint for a complex backend project that i'm working on, and it's fast to talk but slow to finish.
Yes, it can figure out what's wrong in five minutes, but then spend the next forty still going in circles. The tokens fly by, but the actual progress doesn't.
Don't get me wrong, it's still a solid model for the money. Just don't fully judge it by a landing page demo.
2
4
424
Anyone making web AAA rpg with AI that actually have a story?
I feel like all games people make with AI right now are just eye catching for 10 seconds and that’s it. Good graphics and gameplay are only part of making a game, I prefer games that tell a good story, cinematic, and I can immerse myself in that world and enjoy it
Some of my favorite games:
- The Witcher 3
- Cyberpunk 2077
- Death Stranding
AI or no AI, we need more high quality slow burns and less eye grabbing slops
1
137
If DeepSeek can have 95%+ profit margin, let’s assume Anthropic has 99% profit margin
Based on @SemiAnalysis_ subscription usage test result, Claude max offers 58.6x more usage compared to api pricing, that means they still have 41% gross profit margin on Claude max subscription, and that’s assuming all users max out their quota
And the reality is most users don’t even max out their weekly usage
In fact if Anthropic api pricing profit margin is above 98.3%, they can break even on subscriptions
Anthropic is definitely making good money right now
Inference Gross Margins - 87.76% to 96.94%
Assumptions:
97% cached input, 2% uncached input and 1% output tokens
Prefill: 80k uncached input tokens/s/GPU.
Decode: 16k output tokens/s/GPU (250 tokens/s/user)
Token mix: 970M cached input + 20M uncached input + 10M output per billion total tokens.
It takes around 875 GPU seconds (250 for prefill & 625 for decode) to process 20M uncached input + 10M output tokens.
Assuming GPU rental cost is around $3-$6/hr, the GPU system cost per billion tokens is around $0.73-$1.46
1B tokens generate $11.91 in revenue at off-peak prices and $23.82 at peak prices
At off-peak prices, inference margins are between 87.76%-93.88% depending on the GPU rental cost ($3-$6/hr)
At peak prices, inference margins are between 93.88%-96.94% depending on the GPU rental cost ($3-$6/hr)
1
159
At the current climate you can’t just do something silently and expect it to suddenly launch with a bang, it rarely works
You have to consistently build narrative, share while you build publicly, make potential users be the part of the story very early on and keep them engaged
We all learn that the hard way
My friend is crying...... 🥲
Friend spent almost a year building a great product. Scrapped it and rebuilt it several times, and finally got it into great shape, ready to launch.
Today, someone with over 70K followers launched the exact same product he'd been building.....
108