Long intelligence. All in AI.
Hong Kong
Joined October 2021
- Tweets1.5K
- Following1.6K
- Followers170
- Likes2K
I couldn't agree more
We'll be spending a lot more time trying to understand the outputs of language models. A few thoughts, tips & tricks:
Writing. Something I've had success with: Ask your LLM to explain something in ASD-STE100, it's a controlled language specification originally developed for aerospace maintenance documentation. LLMs well-versed in this language and it comes with heavy constraints on clean writing style that I often find a lot more readable. Sometimes I've tried to soften it a bit e.g. ask for "80% of the way to ASD-STE100" because the spec is quite stringent. But even better:
Diagrams / images. Instead of writing, ask your LLM to create a diagram. These can be a lot easier to process, parse, and understand. But even better:
Web pages. Ask for output "in HTML" to get a beautiful, interactive webpage. LLMs are getting really good at frontend and can create beautiful experiences, animations, etc. But even better:
Explainer videos. The output format I am most bullish on is fully custom / bespoke explainer videos generated on any arbitrary topic. Experiment with things like "Create a 3b1b style video explainer on X. Use my ElevenLabs API key for audio narration". (you'd need an API key for the latter or you can ask your LLM to find you decent free alternatives that use your local compute). This is actually starting to work!
In summary:
- As LLMs get better, they will do more and more of the legwork autonomously, and a lot more of our work will rise up the abstractions into oversight and understanding.
- Luckily, LLMs can help here too because as intelligence and code are increasingly abundant, you can ask for large, custom, discardable software artifacts (e.g. web apps, video explainers) that would have never made sense to create before. Push the boundaries here and you'll be surprised.
liskywalker retweeted
An update regarding God of War Laufey's release date.
ALT Since the reveal of God of War Laufey in June, we've been deeply grateful to see so much excitement for Faye's journey through the Everywhen. We don't take for granted the trust placed in our studio to make an extraordinary experience worthy of the passion and dedication of our fans. With this in mind, we've made the decision to delay God of War Laufey's release to May 20th, 2027. We will make the most of each moment until then by putting in the finishing touches and pushing the game to its fullest potential. On behalf of everyone on the team, we appreciate your patience and we'll be back to share more updates soon! Santa Monica Studio
liskywalker retweeted
发布 baoyu-design v2，完整的把 Claude Design、Claude Slides 和 Claude Dashboard 移植过来了，和原版体验几乎一致（除了没有评论功能），可以运行在任何 Agent 任何模型，不再依赖于 Claude Artifacts。
完全开源：github.com/jimliu/baoyu-desi…
AI is reshaping the entire internet. The moats built by tech giants may gradually be broken down by a new generation of AI-powered products.
I look forward to seeing a truly open, open-source operating system that works across different hardware platforms, along with a wave of great applications made possible by AI.
Social media, in particular, is an area where I hope to see real change. I want a platform that genuinely prioritizes content quality. Users should be free to post privately or publicly, while AI handles content evaluation and recommendations, preventing low-quality content from spreading simply because it drives engagement.
No product pushing, livestreaming, spam ads, or telecom scams. Real-identity verification, with open and transparent rules.
Of course, this is just my vision. But I believe that as AI continues to lower the barriers to building products, more and more great ideas will inevitably become reality.
We are truly living through a period of profound change.
AI is reshaping the entire tech and internet landscape. It feels like we're back in the early 2000s, when the industry was a Wild West and everyone was fighting for a place in it.
Except this time, AI is devouring everything at an astonishing pace.
I think it's time to assemble an Avengers team to protect humanity.
🤖 Made with AI
🚀 Introducing DeepInfer, our in-house inference framework, now fully open source!
Powering DeepSeek-V4.1-Flash at scale with high throughput and low latency. We’re sharing our production stack to help the community lower compute costs, reduce energy use, and make AI more affordable.
🔗 github.com/deepseek-ai/DeepI…
1/4
liskywalker retweeted
DeepSeek just open sourced its high-performance, large-scale inference framework DeepInfer in Apache-2.0 license.
github.com/deepseek-ai/DeepI…
Supported features:
- Optimized kernels, model and sampler runtime
- Parallelism: Attention DP / balanced CP, MoE EP + EPLB
- DSpark and confidence-scheduled verification
- SWA bounded replay
- Asynchronous scheduling (zero-overhead batch scheduler)
- Batch-invariant inference
- Deterministic sampling
- Structured output
Computex 2023
Reporter: With AI taking over, has NVIDIA forgotten about PC gamers?
Jensen Huang: RTX was invented for gamers. DLSS was our first major AI focus. Even our first generative AI application was built for gaming.
“You hurt my feeling so deeply.” 💔
— Jensen Huang, NVIDIA CEO
Nvidia denies RTX 5090 discontinuation rumors. Jeff Fisher says shipments will continue despite supply constraints, stressing that GeForce and CUDA laid the groundwork for today’s AI market.
For gamers, the question remains: when can we buy one?
4gamers.com.tw/news/detail/8…
The three most important things in AI are algorithms, compute, and data. You don't need to master all three. A major breakthrough in any one of them is enough to transform the industry—and potentially make you world-famous.
liskywalker retweeted
Introducing earshot
Opus 5.5 can build a whole launch video in code but still produces sloppy sfx.
so we built earshot. it's ears for AI coding agents. they can now hear sounds, judge them and even make their own good sounds.
every sound in this video was made with code.
(sound on!!)
I’m getting increasingly sick of reading long AI-generated articles.
Something that could be explained in a few sentences gets stretched into thousands of words. The whole thing is full of true but pointless statements, with painfully low information density and a sensationalist headline to top it off.
I’m increasingly convinced that good writing should make the core idea clear in the simplest language possible.
The value of writing isn’t in its length, but in how much useful information it actually conveys.
What excites me most about AI is what it puts within reach of individuals and small teams: products that once took far more people and money to build.
More people can now turn an idea into something real that others use. Even the most entrenched platforms feel a little more open to challenge.
It brings back some of the early internet’s scrappy optimism. Trust your judgment. Build something. Put it out into the world.
Technology can change lives—including your own. Keep creating.