HNHacker News
TopNewBestAskShowJobs

bwfan123

1,311 karma · joined December 10, 2024

submissionscomments
bwfan123··on More questions about whether researchers can trust OpenAI with unpublished math
there are also attempts to crowdsource human research directions - like the caltech mathathon challenge : https://mathathonchallenge.com these would help models on the same problems at the expense of the researchers. basically, math researchers are the reverse centaurs but they dont realize it.
bwfan123··on Anthropic researcher says more than 10% chance AI "could kill all humans"
> 10% chance AI "could kill all humans"

Why not 10% chance that it will create enormous prosperity for all ? This is why the average person is increasing pissed at AI in general. That it gets associated with negativity.

bwfan123··on Tao: Open math problems being non-renewably mined by AI
> but at the cost of sustaining the ecosystem

John Naur said the same of coding eons ago [1]. That code is only as healthy as the human mental model of its maintainers. Code is not the artifact, it is the intuitions and insight and mental models of its human maintainers. For a project to sustain, it is vital that the human mental models are kept alive.

[1] Programming as Theory Building https://pages.cs.wisc.edu/~remzi/Naur.pdf

bwfan123··on DeepSeek launching v4.1 flash cheaper and more capable than v4 pro
Same, very impressed with v4 flash. It has the right balance of cost and performance.
bwfan123··on I resigned from Anthropic today
More doomerism. Try to implement a deterministic workflow using agents with the latest models and no humans-in-the-loop, and you will realize what they are really capable of. There is too much unnecessary fear-mongering. All of this is only coming from the 2 AI labs trying to IPO. Not from anyone else.
bwfan123··on Tao: Open math problems being non-renewably mined by AI
It is now clear to me why the AI labs are sponsoring these mathathons: https://mathathonchallenge.com/. They are basically crowdsourcing human researcher data to get access to promising directions possibly later to scoop others.
bwfan123··on Multi-Agents LLM Financial Trading Framework
> The process is quite scientific

This part is lost on many. The value of data is in the theories it confirms or more importantly disconfirms, and defining the trading edge is not easy after accounting for costs. I am wary of black-boxes that produce an edge - not only because I dont know how it works, but also because regimes shift unpredictably, and what works today may stop working tomorrow. That being said, AI can be useful to help automate many routine processes just like any other software.

bwfan123··on Caltech Mathathon – first hackathon ever devoted to research level mathematics
In your final report of this experiment, can you also report the following: How many problems were NOT solved by AI in spite of attempts. What is missing in the debate of AI in math is that only the successful cases are reported. To get a balanced picture of the effectiveness of AI, what is also needed are cases where it failed.
bwfan123··on Nvidia's Jensen Huang says 'AGI has arrived' and congratulates OpenAI
Pretend and extend. Wen rug pull ?
bwfan123··on De-Brainrot Vacations
> Nothing is that novel anymore, and solving day to day problems, more than requiring mental exhaustion, mostly require spending some time at them

I felt this too recently and I linked it to excessive passive information grazing exacerbated by AI. Symptoms of this are being in a daze or mental fog, and not being in touch with the feeling of curiosity. Reading doesnt help since for me, reading is passive process of meme gathering. What helps in my view is effortful creation in the form of writing code, hand writing essays, creating music etc.

bwfan123··on Caltech Mathathon – first hackathon ever devoted to research level mathematics
> Or the big labs trying to use professional mathematitians as (cheap?) Labour for validating LLM outputs?

I am told AGI has been achieved. If so, shouldnt these systems be out and about on their own ? Looking at 1st proof submissions in batch 2 it is clear that fully autonomous AI systems have a long way to go.

AI harnessing human labor with the incentive of 2M in free tokens is the way my skeptic eye sees it, or humans being duped as reverse-centaurs.

bwfan123··on GPU World
> It just requires instructions to continue working until the spec is complete

Try putting an LLM agent in a deterministic workflow without humans in the loop. My experience with this is not encouraging. Getting it to work requires sprinkling some context magic and hoping and praying the LLM does the right thing. More astrology or religion and less science. Great for use cases with humans-in-the-loop, but less than impressive when you need determinism and reliable operation.

bwfan123··on What my dad taught me about AI coding in the 90s
> I feel one need to also have developed the intuition by writing lots of code for reading to be really effective.

This is one of the problems young engineers are going to face who rely excessively on AI to generate code. Their intuitions on what constitutes good code will not sharpen, since they are not exercising the tool that sharpens it which is writing code. Intuitions start fuzzy, and incorrect, and gradually sharpen with precise communication of said intuition in the form of writing code or proofs. Reading alone may delude one into a sense of false mastery where intuitions are actually fuzzy, but one thinks otherwise.

bwfan123··on What my dad taught me about AI coding in the 90s
> See how ridiculous that sounds?

For most of us, writing code is the way to carve out intuition into an artifact. But, I have noticed some people are able to read deeply - and by that I mean, reverse the code to understand the intuition that brought it to life. This is a rare skill and I dont have it, but some do. Not just for code, but also for any book - fiction or non-fiction - some are able to deconstruct the scenarios much better than others, and in that sense understand what they read.

bwfan123··on Warp builds self-improving agents on Claude
> Agents need to handle recurring tasks reliably and effectively

This core problem remains unsolved. The solution presented in the article with Human In The Loop and some skill-magic such as "Write principles, not rules etc." is unsatisfactory because it offers no guarantees whatsoever. I find it difficult to harness agents into deterministic workflows which need to produce reliable outcomes.

bwfan123··on Bhartrhari's Paradox
> Isn't this just a proof that there are many unnamed things, but no unnameable ones?

Where is it a proof that there are many unnamed things ? I could only see it as an argument that there are no "unnameable" things.

bwfan123··on Bhartrhari's Paradox
Names are like variables in a function. you can name variables anything you want from a human understanding point of view (final cause), but the compiler doesnt care about that. The compiler only cares about the efficient cause of that variable in the sense of what it represents (stack/heap etc).
bwfan123··on OpenAI Jalapeño: Better than Nvidia Blackwell
> some of the statements seem like just straight up lies

Why the angst ? I suspect this announcement punctured a lot of people's bubbles, and many are in disbelief and denial and hence the emotional reaction seen here. That a company which never designed chips could suddenly leapfrog the best in the industry. What many forget is that openAI and anthropic are in a unique position to own the end-user experience, and that provides them a distinct advantage. But, making announcements and actually delivering are two different things, and it remains to be seen if these are actually viable. In any case, it gives openai leverage over their vendors.

bwfan123··on The state of AI in 2026: On the road to ROI
> Please hire more McKinsey consultants who say they’re AI experts

You have to recognize that a vast majority of enterprises fall in the category of "Dumb IT". To them, IT is a cost-center and overhead as their main line of business is something else. And, they look up to the likes of McKinsey for wisdom.

bwfan123··on Could AIs Become Conscious?
> At some point mimicking a thing becomes the thing.

You can have a long conversation with AI about pain and suffering, but you have felt it and AI hasnt. The difference is that, for us humans, words are grounded by something other than other-words, words are grounded by experience. Meaning for us is operational in the real world in the form of perception, feeling and emotions, and movement, and words symbolize said meaning. Meaning for LLMs is something in the concept-space that is detached from any worlds. I suspect there should be a test which can tease apart the ungroundedness of the LLMs just from having a conversation with it.

bwfan123··on The Mojo language (by Modular, now Qualcomm) is now open-source
> expert C++ programmers as what they imagine a good C++ will look like

Mojo looks nice but with coding agents the landscape has shifted at least in my view. There is no need for DSLs and abstractions that reduce human cognitive load. So, I still struggle with the usecase for Mojo. If you want close to metal performance, code in something that touches metal (c,c++), and doesnt abstract you away from it or require unsafe primitives. As a programmer who cares for performance, you want the least separation between your tools and the hw in the sense that you can visualize precisely what the hw is doing with no abstractions to hide it.

bwfan123··on GPT-5.6 Sol Pricing Cut by 50%
> The AI bubble is just waiting for the first player to scream mercy and cut capex as they simply can’t afford to throw more cash on the burning pile. That will be the trigger that implodes this bubble

Deepseek v4 and Kimi k3 have tightened the screws on the frontier models. With open-weights, anyone can host these for cost of compute. So, there is zero leverage left for the frontier labs.

What implodes it in my view is that enterprise adoption will stall. Enterprises are struggling to actually use these things in real workflows outside of coding and support.

bwfan123··on The AI Credit Resale Economy
> Cyberpunk dystopia vibes

right. Abstractions taken to the max. When tech solves problems that only 0.001% care about. NFT smelt similar.

bwfan123··on Why does Opus 5 feel worse to work with?
> CLAUDE.md is mostly powerless against the reinforcement learned crap.

When you dont know the cause, you dont have a fix. Thats the biggest issue i have with all of AI is that we dont know how it works, and yet we think it will be great ! This is more like a religious belief than a scientific one. There is no causal model of how it works, there is no theory. And the temerity to call it intelligence is annoying.

bwfan123··on When Genius Fails: The Intellectual Arrogance of the AI Labs
AI is just another tool in the arsenal of tools for both developers (Linus [1]) and movie makers (Ben Affleck [2]). These are expert practitioners in their fields intelligently opining on AI, and they are promoting it not as a magic machine, but as another tool in the toolbox.

[1] https://www.youtube.com/watch?v=3NSSGt9bZag

[2] https://www.youtube.com/watch?v=O-2OsvVJC0s

bwfan123··on Retire the Abstractions
"The question is whether letting it sit higher is a difference in degree or in kind, and we don't think it's in kind"

I was scratching my head reading it. The cognitive load was high with cryptic sentences like the above. AI-slop or not, there is no excuse for dense prose particularly if it is intended for a human reader.

bwfan123··on What sort of maths are LLMs good at?
From the tweet:

> Still not sure what that means, but some analytic number theorists seem excited

ie, I prompted AI and it put out a giant pile of tokens. I dont know what it means, but I hope someone gets excited. Mathematicians are now the priests and shamans chanting incantations and taking the holy blessings from the AI gods.

bwfan123··on Nvidia's Risky Business
> overzealous peddlers and hacks

They are the marketing wing of the AI ecosystem. Their recent article on how SpaceX would drive 500B in data-center revenue was ludicrous-mode. Lets revisit this in a few years.

bwfan123··on Learning more about Claude's mathematical capabilities
> Jarred prompted Claude to try again, and it spent a day and a half coordinating about 60 Claude subagents, which this time went much deeper: between them, they ran 2,400 shell commands and wrote hundreds of Python scripts.1

If there is anything to learn from the history of science, it is that breakthroughs happen via better or new theory and not by brute-force compute [1].

[1] https://arxiv.org/pdf/2607.27794

bwfan123··on Muse Glimmer: 30B-parameter model optimized for always-on local agent workflows
Next step: Burn the weights of these local models into an asic that ships cheap on a laptop (AMD/taalas looking at you), and I will be a happy camper. Make it pluggable so I can select a model I want. I use qwen3.5-coder currently on my laptop, and while it works well enough for me, it is somewhat slow processing tokens.

I would hazard a guess that fast small models with a smart agent harness can do quite well compared to large models which cant be run locally.

← PreviousPage 2 of 18Next →