Hacker Newsnew | past | comments | ask | show | jobs | submit | 2gremlin181's commentslogin

I do not forsee GoogleClaw, MetaClaw, and AppleClaw all playing well with each other. Everyone will have their own walled garden and we will be no better off than we are now.

I knew I had to add GLaDOS as soon as I saw this. Unfortunately, while testing my PR I realized there’s no support for Linux. Hopefully someone smarter than me can get that added sooner rather than later.


I am using piper-tts ( https://github.com/OHF-Voice/piper1-gpl ) with these voice files for GLaDOS: https://huggingface.co/rokeya71/VITS-Piper-GlaDOS-en-onnx/tr...

It is not perfect, but quite sufficient for simple system messages.


Is there any reason why one would use an ad-filled ChatGPT over any alternative or open-source LLM providers? I feel like things have stagnated from a model perspective for simple queries one might ask ChatGPT. The key differentiators for it being their user intent understanding, web search tooling, and deep research/thinking mode, all of which are much smaller moats compared to training an LLM.


I think most free tier users will stick with chatgpt given its brand stickiness and lack of obstacles (disposable login page). If you can run your own llm models you’re definitely not the target demographic


What makes this any more personalized than Google Search ads?


This is working a lot like Meta/Facebook where they have got immense data about your interests.

Google Search OTOH has been using broad matched queries and is deciding which keywords to show your Ads.

I heard from many people that they don't like this approach of Google Search Ads now. As they are blowing up more money for useless keywords they didn't want to target. The only option they have is to add negative keywords - that mostly happens after the money is spent on junk keywords.


Senior Data Scientist with 5+ years of experience building and scaling ML, analytics, and GenAI systems. Expertise in large-scale data analysis, model evaluation, and experimentation. Proficient in Python, SQL, Spark, and statistics, with a Computer Engineering foundation.

  Location: San Francisco, CA  
  Remote: Yes  
  Willing to relocate: No  
  Technologies:  
    * Programming Languages: Python, SQL (Presto, Postgres, Spark), R, PHP/Hack, Bash  
    * Frameworks: PyData (Pandas, NumPy, scikit-learn), PyTorch/Lightning, TensorFlow/Keras, PySpark, LangChain, XGBoost  
    * Tools: AWS, GCP, Spark, Gurobi, Kubernetes, Docker, Linux, Git, Tableau, OpenAI API  
  Résumé/CV: https://kavi.sh/assets/resume/KavishHukmani_resume.pdf  
  Email: khukmani at gmail.com  
  GitHub: https://github.com/DoubleGremlin181  
  Website: https://kavi.sh/  
  LinkedIn: https://www.linkedin.com/in/kavish-hukmani/


I acquired a new domain last year: https://kavi.sh/


Search results is by far the biggest one. Now after 3-4 relevant results it just shows shorts and other recommended content


What’s even the point of this? Is showing random videos really driving up engagement more than surfacing the video I want to watch? This was shocking to me when I first came across it


Yes. They know that most humans typically have poor impulse control, and are easily pulled off task and will fall into an addicting and lucrative loop. Makes perfect sense to show random unrelated shit.


It's been like this for a long time so it must be working. My guess is that fewer choices = easier to choose a video.


But it's not fewer choices, it's worse choices, or more distracting choices.


The article is glaring over a key point- Valve often sold the model at a 20% discount for $320. Clearly this is because the amount spent by a Steam Deck owner would make up the loss in BoM. If they believed they could continue to do so, they would.


Genuinely curious, what are some use cases that you require live Twitter data in your LLM for?


The topic of this HN thread: security, which is ever-evolving.


IMO this is a smart move. A lot of these next-gen dev tools are genuinely great, but the ecosystem is fragmented and the subscriptions add up quickly. If Cursor aquires a few more, like Warp or Linear, they can become a very compelling all-in-one dev platform.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: