LewisOrton

Thanks. i thought about this but eventually figured out it does not really work in real environment, speaking it out loud can be a bit awkward (and not efficient). if there are enough requests i will surely add it but in a responsible way, not sure if voice models today are good enough for many languages

If I just ask questions and do simple research without proactive local works, I'll just use the aistudio, it's probably 10x better than the Gemini app. But other than that I have my own fork based on dsh (nothing fancy I just added a few tools & fixed the thinking level param issue) I rarely use it though. Computer use is bad and i mostly use codex for that

i dont think they've nerfed it did you use api or via subscription(which might make sense) api has been pretty stable for me.

personally i like 4 Argon, it's very clever for me in a very different way comparing to the other two (especially via api) but when it comes to the google products.. sigh. nobody who's actually working on these products really cares about them in google now (maybe notebooklm is the rare exception), they're either not dogfooding themselves or somehow intentionally poisoning the experience on purpose, or simply inqualified. poor gemini is suffering in agy i can tell, it struggles so much and wont even use simple tools properly. if you put it in any good 3rd party harness out there it's flexing way better than agy and this tells a lot.

tbh i might be one of the few who actually likes gemini since 3. it wasnt as good at coding compared to other two but it always wrote the best, and talked the best, providing very good conversation experience. it works incredibly well in my customer service agent harness and my users have been very satisfied. Now Gemini 4 is incredible, but the same old issue: a powerful model, in the worst official harness

deepmind should get rid of those antigravity ppl they're a joke and sabotaging the model's reputation.

ok, I think this overlooks two important things:

1. what we currently perceive as a lack of capability in a particular domain often comes down to the model simply not having received targeted RL training for that domain (yet). Look at Opus 5.5 in motion graphics: after its first round of specialized RL, it was already producing work that surpassed the average professional designers(and it's using vanilla javascript vs human with After Effects, let that sink in). It doesn't need to beat the top 1% of designers to fundamentally disrupt the profession. Most professionals aren't the top 1%.

2 people rarely hold agents to the same standards and workflows they apply to human professionals. A human designer gets to iterate and refine until the work meets expectations. Yet ppl often judge AI based on oneshot or fewshot attempts (these are just "wishes") and treat whatever comes out as the ceiling of its capabilities. Why shouldn't an agent be allowed to iterate against the same acceptance criteria until the result is good enough?

that's before considering domain-specific skills& loops, which can dramatically improve performance in specialized tasks.

The relevant comparison isn't an experienced professional vs LLM given a prompt five mins ago,it's a professional vs agentic system w access to the same project context(including tools and feedback, and opportunities to iterate)

I dont know if i fully agree to it. ppl are still trying to find the aspects that AI do not handle (or do very well) yet, to define the irreplaceable value of human (humanity?), without realizing that these abilities/capabilities are within the expanding range of AI soon if not tomorrow. Why would one assume that (99.99% of) human can solve the issues that AI can not?

should we stop treating the idea that humans must possess some capability AI can never replace as the starting point for discussions about humanity's future?

I'm inclined to think we should.

Preserving human agency is a moral and institutional choice, not something that needs to be justified by human cognitive superiority

Do androids feed electric ducks

btw i found it works very well as white noise

it's a desktop app so i dont see any reason not to use avif, thanks!

in my app i was using jpeg but then switched to webp two years ago, they're for generating/displaying thumbnails for images and videos. and now avif seems like a better fit? anything that you're not happy about? considering avif now, appreciate any ideas

it's actually more usefull than it looks. it is not much useful on its own but it can play a big role in a complex workflow or custom harness.

i used luna 5.6 as the ranker/filter to remove useless items from upstream rag(mixing embedding, regex etc) and it did a good job but took 5-10 seconds each time. now less than 1 second and no observable precision loss. hope it helps

a unified solution for both old and new problems! such a nice idea. i liked gmail's "+" variant but yours are way cooler

exactly, it sounds more like a trap with the facade of benefit than passkey

How did I not know this until today? it's so beautiful, thanks so much!

a few opinions: the detection of the mouse hovering isn't forgiving enough and sometimes it gets a bit annoying. try to use bbox with some padding wherever possible, and add some visual hints when hovering over the notes (slight scaling up etc)

There is plenty of space for improvement like others have said here and i think it just needs a bit more time.

llms are great at languages but you dont just ask it to "translate from english to other supported languagues" it generates things that are hard to understand. when we talk about translation in real world we're actually talking about is a fundamental re-intepretation(transcreation) of the source text through target language's local, native mindset, and you should do the same with llms.

So could Astra, it uses pythons but just for assembling the final exe and the bytes are in the script. It makes simple stuff like a calculator and you can even asked to change/fix anything not working correctly. also I've tested astra can write pixels to a tiff to get an image directly, without using any other tools

it's still vastly underestimated how much these models can do and what would happen once these domains are scaled up

after a “not sure,” how often does reprhasing get a correct answer on the next try/do users give up? retries could make a simple lookup surprisingly slow?

i was running 300-500 codex agents simutaneously from time to time these few days on my little 32gb ram mbp they worked pretty well.

It works when the time comes for you (you realize you do need the army for your job, dont try to use swarms for using swarms only, don't create the situation on purpose)

my pattern: one chat as the main orchestrator that distributed (predefined, scoped) jobs to independent chats (created by the main chat), and then they autonomously spawned fresh sub agents (research, collecting context from various inputs, edit, test, audit etc ) following required workflow to finish each part independently. not by autonomously, recursively delegating sub agents, that's a failed mode and will be a disaster.

the concept that worked for me here is to manage the boundaries by considering reducing agents' cognitive load

You must be someone that needs something, then you will learn why chatgpt is not always the solution, despite it can do things, but there is so much friction in many tasks, because you need to take time to navigate around threads, type, etc. and in real world these actions can be awkward when the situation comes.

It's used extensively in my own agentic products. basically as a filter for various purposes, one example, to reduce the main agent's cognitive load.

there was a dynamic context loader (mixer) that combined results from embedding top k, regex filters etc and then there was a sub agent as a ranker to filter out the useless ones not matching the queries (wrong results from regex, and useless ones from top k etc). it was the job of a 5.6 luna(which worked very well ngl) but then replaced by jev.

Saved 5-10s in each round, no precision loss, and saved the bill a bit more. net win for me

Before I knew it, ten minutes had gone! good job

ngl gemini is helping to provide (real helpful, not the ones giving runarounds) customer services in one of my own products.

i benchmarked 20+ models in real worlds and gemini families are the best (highest customer satisfaction)

Gemini models are in fact very good. Google's harness(es) for Gemini are nightmares, terrible