HN user
Just ran evals on a few detectors, and clark-browser passed live stealth checks on Cloudflare smoke, SannySoft, BrowserLeaks Client Hints, and WebGL.
Results + new binary https://github.com/clark-labs-inc/clark-browser
EXPENSIVE ._.
clark-browser is c++ level overwrite of core chromium parts
I also cancelled my subscription.The $20 Pro plan has become completely unusable for any real work. What is especially frustrating is that Claude Chat and Claude Code now share the exact same usage limits — it makes zero sense from a product standpoint when the workflows are so different. Even the $200 Max plan got heavily nerfed. What used to easily last me a full week (or more) of solid daily use now burns out in just a few days. Combined with the quality drop and unpredictable token consumption, it simply stopped being worth it.
* fine tuned Qwen-7B
That could totally power next generation of green-screen techs. Generative actors may not find favorable response in the audiences; but SFX, decor, extras, environments that react to actors' actions - amazing potential.
needed to find a kids’ orthodontist. made a tiny voice agent: feed it numbers, it calls, asks about price/availability/insurance, logs the gist.
it kind of worked. the magic was the smallest UI around it:
- timeline of dials + retries
- "call me back" flags
- when it tried, who picked up
- short summaries with links to the raw transcript
once i could see the behavior, it stopped feeling spooky and started feeling useful.
so yeah, copilots are cool, but i want HUDs: quiet most of the time, glanceable, easy to interrupt, receipts for every action.
props to OP for the screenshots and payloads—that’s how you do it. If any IDE wants trust, they know the recipe - make telemetry optin by default and provide a real kill switch.
Engineers: always negotiate for higher base salaries. In the vast majority of cases—especially during acquihires—your equity will be worth little or nothing. Founders and VCs still get paid; employees rarely do.
Don't just accept promises. Ask for the 409A valuation, liquidation preferences, and pay bands. If a company won’t provide transparency, that’s your signal.
Equity is a lottery ticket. Salary is money in the bank.
exactly for this made vortle.com no login, no internet needed, no data collected, no score - just you and the game to de-stress and calm down
For most tasks, GPT-4o/o3-mini are already great, and cheaper.
What is the real-world use case where GPT-4.5? Anyone actually seeing a game-changing difference? Please share your prompts!
Nice! Built something similar in Rust that supports local and remote repos: https://crates.io/crates/r2md
made this super easy to use tool https://github.com/skirdey-inflection/r2md
I've been comparing R1 to O1 and O1-pro, mostly in coding, refactoring and understanding of open source code.
I can say that R1 is on par with O1. But not as deep and capable as O1-pro. R1 is also a lot more useful than Sonnete. I actually haven't used Sonnete in awhile.
R1 is also comparable to the Gemini Flash Thinking 2.0 model, but in coding I feel like R1 gives me code that works without too much tweaking.
I often give entire open-source project's codebase (or big part of code) to all of them and ask the same question - like add a plugin, or fix xyz, etc. O1-pro is still a clear and expensive winner. But if I were to choose the second best, I would say R1.
Is anyone fine-tuning llama to write Q#? I feel LLMs can be a helpful tool in learning how to code quantum systems.
Google should just offer llama3 405b, maybe slightly fine tuned. Geminis are unusable.
does the crew get paid overtime if they were to stay until Feb?
I've been diligently trying to use Gemini 1.5 Pro, and it is not even on the level of Llama3-70B. I really hope Gemini improves, even if it gets reduced context length.
I've been testing Ultra model today and comparing it to open source Mixtral MOE on HF chat. Gemini Ultra lost in every instance to the free open source model, including code generation. I think the tasks I do with the help of LLMs are common, and Gemini Ultra is unusable at the moment.
Gemini refuses to answer or perform even on the simple prompts.
I hope Google team can make it better, but at the moment, for my light coding and text analysis use cases it is not worth $19.99
took one AG to try to cancel their subscription :D
Claude2 summary: Google has amazing resources and access, but does things "the Google way" - they rebuilt the Socratic product from scratch using Google's own stacks and to their standards.
Simple things done repeatedly can feel magical at Google's scale - like recalculating signals across the entire internet to improve Search. But much improvement comes from manual analysis and labeling data.
Surprising problems get tackled if the right teams are interested - like developing a math image recognition API from scratch in 6 months. But most products face many hurdles to launch.
There is an ever-shifting web of goals and efforts. Politics and frequent re-orgs can derail projects. Smart people argue rather than align.
Technical debt is real, but so is process debt. Layers of reviews and requirements accumulate over time. Top-heavy teams with lots of senior people can cause gridlock. More doers than thinkers are needed.
To drive something big, you must relentlessly sell the vision and get the right leaders on board.
Many acquisitions fail. The Socratic founders left, and some goals weren't achieved, but parts of the product grew significantly.
Overall, amazing things are possible at Google if you navigate politics, rally the right support, and play the long game. But it's challenging due to complex processes, shifting priorities, and ingrained ways of doing things.
claude2 tldr: Here are a few key points I gathered from the article:
The article explores optimizing LoRA hyperparameter settings for finetuning large language models. The goal is to maximize performance while minimizing memory usage and training time.
The base model used is Llama-2 7B. Experiments compare default LoRA, QLoRA (4-bit quantized), AdamW, SGD optimizers, and different choices for rank r and alpha hyperparameters.
Key findings:
QLoRA provides substantial memory savings (6GB less than default LoRA) at the cost of slower training. Performance impact is minor.
AdamW vs SGD makes little difference in memory or performance.
Increasing training iterations from 50k to 100k hurts performance, likely because the Alpaca dataset lacks diversity.
Tuning rank r and alpha is most impactful. Good rule of thumb is to set alpha=2*r. Best model uses r=256, alpha=512. Improves over base model on most tasks, except arithmetic.
The optimized LoRA model was submitted to the NeurIPS efficiency challenge and showed improvements on several benchmarks compared to the base Llama-2 model.
Takeaways are practical tips for tuning LoRA hyperparameters and trading off memory, compute, and model performance.
we all seen it in the show, old news
just curious and trying to understand better what text autocompletion is capable of
makes sense
good to know, thank you for the note! maybe there can be a prompt that can investigate molecular properties?
as someone who don't know anything about anything I've been bugging GPT4 with the questions about this material, if you are curious to see the conversation https://chat.openai.com/share/50bceb29-7042-40aa-bc13-a97443...
please, issue corrections in the comments if you have a second
there is an alternative that is production-grade - deepset haystack https://haystack.deepset.ai/
p.s. i am contributor so there could be bias
it will quickly turn into an e-graveyard
I am building something similar, it has documentation from azure, aws, and lots of slack/discord threads of software projects are also searchable, check it out https://www.kwq.ai
it even gets real time indexing from slack of aws deep java library, and from discord of deepset haystack project