nopinsight

7 ms·

61.
by nopinsight 1y ago
What would you say if the IMO Gold Medal models from DeepMind and OpenAI turn out to be generalizable to other domains including those with hard-to-verify reward signals? Hint: Researchers from both companies said publicly they employ gener
62.
by nopinsight 1y ago
Sounds good in theory but I'd be bored to death in a month, at most two. Traveling the world...maybe good for a few more months and that's it. Wouldn't you yearn for any more impact given how much that amount of resource cou
63.
by nopinsight 1y ago
impact != story points
64.
Tech Philosophy and AI Opportunity(stratechery.com)
1 points by nopinsight 1y ago|0 comments
65.
by nopinsight 1y ago
My thesis is that this could lead to a booming market for “pink-collar” service jobs. A significant latent demand exists for more and better services in developed countries. For instance, upper-middle-class and middle-class individuals in c
66.
by nopinsight 1y ago
Wondering if you could add “Flux 1.1 Pro Ultra” to the site? It’s supposed to be the best among the Flux family of models, and far better than Flux Dev (3rd among your current candidates) at prompt adherence. Adding it would also provide a
67.
by nopinsight 1y ago
...which is why top LLM providers' web apps like ChatGPT, Claude.ai, Gemini try to nudge you to connect with Google Drive, and where appropriate, GitHub Repos. They also allow the user/dev to provide feedback to revise the results
68.
by nopinsight 1y ago
Words are more than just symbols; they represent concepts and patterns we observe in the world, in our society, and inside ourselves. Translation is only possible because we are all humans and have experienced broadly similar concepts, but
69.
by nopinsight 1y ago
>> ...median urban household income... You seem to miss the word 'disposable' when you cited the chart and table in [0] you linked to. China's nominal GDP per capita is already on par with Malaysia's. https:&#x2
70.
Robin: A multi-agent system for automating scientific discovery(arxiv.org)
151 points by nopinsight 1y ago|20 comments
71.
by nopinsight 1y ago
Au contraire, reasoning LLMs free me from mundane cognitive tasks, e.g. selecting websites to read from among search results, assessing level of evidence in papers before reading them in more detail, routine coding, and many other repetitiv
72.
by nopinsight 1y ago
Training data, esp interaction data from agentic coding tools, are important for that. See also: Windsurf acquisition.
73.
by nopinsight 1y ago
There’s a recent article with experiments suggesting LLMs are better at bug fixing than coding, iirc. It’s from a company with a relevant product though.
74.
by nopinsight 1y ago
With Agentic RL training and sufficient data, AI operating at the level of average senior engineers should become plausible in a couple to a few years. Top-tier engineers who integrate a deep understanding of business and user needs into
75.
by nopinsight 1y ago
> 4.5/o3 doesn't seem hugely more intelligent then 3.0 I disagree with 3.0, but perhaps that feels true for 4.0 or even 3.5 for some queries. The reason is that when LLMs are asked questions whose answers can be interpolated
76.
by nopinsight 1y ago
She appears to be a very good product developer. However, unless the hypothetical VC knows more about their upcoming plan, it doesn't look like a fundable business with a strong moat and the potential to become a unicorn or larger busi
77.
by nopinsight 1y ago
Livebench.ai actually suggests the new version is better on most things. https://livebench.ai/#/
78.
by nopinsight 1y ago
Current models are quite far away from human-level physical reasoning (paper below). An upcoming version of models trained on world simulation will probably do much better. PHYBench: Holistic Evaluation of Physical Perception and Reasoning
79.
by nopinsight 1y ago
"Can AI meaningfully help with bioweapons creation? On our new Virology Capabilities Test (VCT), frontier LLMs display the expert-level tacit knowledge needed to troubleshoot wet lab protocols." OpenAI’s o3 now outperforms 94% of
80.
AIs Are Disseminating Expert-Level Virology Skills(ai-frontiers.org)
3 points by nopinsight 1y ago|1 comments
81.
by nopinsight 1y ago
> software engineering skills are easier to train than other skills. I think the author meant it's easier to train (reasoning) LLM models on [coding] skills than most other tasks. I agree on that. Data abundance, near-immediate feed
82.
Forecaster reacts: METR's bombshell paper about AI acceleration(peterwildeford.substack.com)
49 points by nopinsight 1y ago|74 comments
83.
by nopinsight 1y ago
The SWE-bench verified score for frontier LLMs will probably reach/surpass 90% by the end of the year. Agentic AI will learn to complete a larger and larger chunk of the practical software development process without much human input.
84.
by nopinsight 1y ago
One way to think of this is as the Baumol effect* within software development. Expert humans are still quite a bit better than LLMs at nuanced requirements understanding and architectural design for now. Actual coding will increasingly be
85.
by nopinsight 1y ago
https://news.ycombinator.com/item?id=43743198
86.
by nopinsight 1y ago
Verbatim quote: “The difference between me and an average Joe in strength technique and speed was still less than the difference between me and the Olympic level athletes that I occasionally competed (or worse, trained) with.” I’m not disp
87.
by nopinsight 1y ago
I'm curious. Do you mean it figuratively? I ask Claude Sonnet 3.7 Extended Thinking since it's usually reliable and the stats for running strongly suggest that for most competitions, the top percentile is closer to world-class ath
88.
by nopinsight 2y ago
The reward it gets from the reinforcement learning (RL) process probably didn’t include a sufficiently strong weight on being truthful. Reward engineering for RL might be the most important area of research in AI now.
89.
by nopinsight 2y ago
Many of them likely won’t switch immediately. They could also try to keep them with sweet offers, like generous usage quotas, early access to the latest models, etc. Once sufficient data is gathered, the next generation models will be among
90.
by nopinsight 2y ago
OpenAI‘s early investment in Cursor was a masterstroke. Acquiring Windsurf would be another. Next advances in coding AI depend on real-world coding data, esp how professional developers use agentic AI for coding + other tasks. RL works well

More ›