HN user
You can't please everyone on the internet, if he had posted this on Arxiv then someone else will be complaining that Arxiv is for humans to publish and that llm output should be social media post instead. Also, the proof fits in a tweet. So why blow it up.
that's because they said they'll silently fallback in case of model training. Then they backtracked, and said they wont silently fallback, but how will anyone ever know?
My daughter suffers from Autism Level 3 and I worked with my daughter's speech therapist and OT to build tejutalks.com which helps nonverbal children communicate with the help of AI.
It's being used by around 10 kids for the last 3 months and it's really helped some of them to demonstrate that they know a lot more than they can say.
It uses Openai models underneath, although initially I started with local inference using MLX. If anyone is curious feel free to reach out.
Hi, thanks for the writeup, I wonder for the auth problem what you think about rego and OPA type solutions and their place in world in comparison to generated guards?
When a qualifying noun is absent , then priors means prior beliefs.
You do realize that djmips is Demis' account?
I'd be interested to learn what kind of internal tooling are you improving ?
... the person next to him is probably also a combatant
Absolutely, they could even be future combatants even if they are not now. That's why killing schoolgirls in Iran, reporters in Lebanon, etc. is justified, they are all potential terrorists. It definitely can not be proven otherwise that they are not. Why take a chance? /s
Building VizPy, a prompt optimizer we've been working on for a while now.
The problem it's solving is one most people building with LLMs know well. Your prompt fails on some inputs, you don't really know why, and you end up just tweaking and re-running until something sticks. We kept hitting this ourselves and it felt like there had to be a better way than guessing.
What we figured out after a lot of research: prompt failures almost always follow a pattern. The model isn't failing randomly, it's consistently failing on a particular type of input or reasoning step. VizPy finds that pattern, distills it into a plain English rule you can actually read, and then rewrites your prompt around it. You also get the rule itself so you can review it, tweak it, or just drop it into your existing prompt directly. DSPy-compatible, no pipeline rewrite needed.
We have compared it extensively against baselines such as GEPA on benchmarks like BBH, HotPotQA, GPQA Diamond, and GDPR-Bench and VizPy wins on all of them. We'll have more benchmarks on cyber-security and chip-design coming out soon.
Free to try, 10 runs no card: https://vizpy.vizops.ai/
Anthropic released a C compiler this week that was built autonomously. We mined the entire repo for clues to reproduce the scaffolding they used and open sourced it.
FYI, It wasn't taken the money and leave, a lot of them got absorbed into GDM.
Source: I was in GDM when character was acquired.
+1 to this, my mom died because of COVID in India 1 months after she left US after visiting me, and I still feel guilty that I didnt insist on getting her the vaccine before she left for India, and then at the time of her death India was locked down so no flights and I wasn't even next to her. It's been 4 years but every so often I think about this. I blame myself less now after some therapy, but If you didn't try all that you could you'd probably feel guilty like me.
Yeah I don't understand why the whole thread has been so hostile against a very reasonable/useful observation that you made. If there was a way to prompt commenters to be less snarky on HN that'd be a vast improvement.
The brits were also running the opium shops. So if you think country A selling opiods in country B and then going to war so that country B can not stop the sale of opiods is totally okay, then that is truly very different from my model of good behavior.
Have you seen YouTube shorts and Instagram reels. Lol
I don't know if you are just ignorant about history and unwilling to Google, or if you are making the point that of course British did not force feed opium to the people.
What is very well established is that the british fought a war , literally called the opium war by Western historians themselves with the main objective of keeping their opium distribution into China open after the emperor banned it
Their action was akin to if some majority owner of Purdue pharma invades US and forces US government to "keep the oxy market open" while letting "people make their own decision".
The hidden variable is the previa marketing budget. They have budget right now for a billboard and for online ads at the same time and they are focusing on your geography
A few months ago I read about a lady who can smell Alzheimer's and at first she too was not taken seriously. I can't think of how RF can affect human body but I wouldn't dismiss their reports completely.
The real fun will be when there is enough of this chemical in our waste runoff that cattle and wildlife gives up eating
Details are in the actual paper, and its linked prominently but it take a few clicks to get there.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10960378/
There's even a video abstract!
I feel this, I also still dont understand RACI , like how can responsibility and accountability be truly separate? It just seems wrong
Many years ago, I did that when I had a large paper reviewing load during my phd. My solution was simply to purchase an app called SayIt for like a dollar that read the pdf to me, worked really well.
Nowadays I often pass the pdf through LLMs to get personalize (expand on jargon or contract the verbiage) and then read them. That gives me a better return on time spent.
Fyi, the pinsker inequality bounds KL divergence in terms of Total variation distance and TVD is like infinity norm on the difference between probability distribution, and sum of absolute differences is the L1 norm, and L1 and L_infty are also related.
tl;dr is to not worry about the mathematical details, if it works it works.
This discovery (if true) is as improbable and wonderful as luis pasteur and bacteria or fleming and penicilin
You are forgetting about developers who may want to develop on top of something stable and with long term support. That's a big market.
If you are an engg in a big company you could just email the relevant mailing list / slack channel. And somebody highly knowledgeable will tell you a lot of what you need to know .
Median is 50percentile. 50% of employees are making less than 80k.
You'd need a scatter plot of pay vs experience in this cohort to really say that the demands are unreasonable.
We dont know how much buearocracy/middle mgmt there is in these companies, and how much they make, and how much value they contribute, vs extract.
I dont see how any lay person can pass judgement about whether the strike is fair or not without access to a lot more facts than just ceo pay or median pay.
Not saying that the strike is justified, just that the over confident tone that its not is unwarranted.
I am surprised that people aren't using google colab pro/pro+ in this context. You basically get access to multiple A100 for $10/month and with some simple javascript tricks, you can get a session to last for 24hrs at least.
Pro+ is more expensive at $50/mo but it allows for more simplified background execution. if you are only just getting started and don't expect to be training for multiple months, then colab or other cloud-notebook providers are really great to start.
Hi OP here. Yes you are right about the asymmetrical nature of the penalties. Actually I do handle this case by using an asymmetric `shrinkage` on the error. See the function called shrinkage in the notebook.
Good day :)
Hi OP here.
To answer why PID, basically somebody I follow asked this question on twitter, and I thought yeah why not seems like a reasonable thing to try :)
Actually I do conclude that PID is not quite the right thing for this problem. For me the learnings from making a PID sort of work for this problem were:
1. must use the right error function. like frequency not time. 2. must use shrinkage on the error to handle discrete number of server. 3. have to run controller at a multiple of server delay to avoid perturbations.
Also I discuss the basic assumptions that a PID controller makes that are suboptimal in the video.