"Find and deanonymize online hate speech accounts. Make some mistakes. (that are politically convenient)"

Are there commies on nostr? Fiatjaf needs to kick them off

The majority of people are as smart as an LLM. They don't actually think, their reasoning is literally just predicting the next token. They genuinely can not form a logical if A, then B, then C chain. During most of their daily life they can present the illusion of logical reasoning, but like an LLM, it falls apart if you ever get past small talk. They can not catch logical contradictions in truth coded phrasing. That is the most basic test that is used to separate a reasoning LLM model from a pure next word generating one. They lack the knowledge of an LLMs training data, and if they are arrogant they will hallucinate complete bullshit facts without looking it up, far worse than opus on low. A lot of them also have a dogshit context window. They are unable to keep the relevant facts in a given conversation in their head. Most of media is staffed by these people. They write news articles indistinguishable from an LLM for a living. When these people see AI, they rightfully see that it does everything they do better than them, but since they were conditioned to believe that their obedience was a sign of intelligence, they underestimate the people more intelligent then them who are actually the ones getting this tech to produce the amazing results it does. A very basic test, which lots of people are trying every day, is to let an agent loose and having it try to make money online. As far as I know, no one has successfully gotten a positive return from this. Think about that. Not even the smartest agents can do any form of work worth as much as their own token price, without a human to lead it. This is much like these midwit journalists. They can work at multi billion dollar media companies and produce content that makes millions, but almost none of them would make even minimum wage if they tried doing anything on their own. Of course they feel threatened by AI, while the smart people of every income bracket enjoy its output.

When you get snitched on by your bitch neighbor to the town for grass too long

Hate these companies that make up "feature" wording to denote something that should be a stat. Why the fuck your screwdriver labelled as "extended reach"? Just say it's 10 inches long.

You can drive around ice cream, you can't drive around a food truck. Theres plenty of food trucks, they are just parked somewhere all day bc you can't drive with hot oil sloshing in the fryer.

Buddy no one considers y'all an enemy (even trump). He just has a bully personality and likes bullying the weak.

Quick benchmark today comparing Astra 6, Fable 5.1 and Opus 5: I have a build plan in my project which has step by step MDs to get to the next milestone. I had GLM 5.3 build 3 of the steps that could be built in parallel. The benchmark was for these 3 models to review the code. They were told to review the code for those 3 commits, not the whole project. Opus 5 was run with Claude code on the $20 plan, the other 2 were used through opencode with API billing. I had all of them review the changes with the same prompt, then I gave them all the outputs from the other 2 agents and asked them to rank themselves and the other 2. Here is the results Astra ranking: A, O, F Opus ranking: O/A tie, F Fable ranking: O, A, F They were not told the names of the other agents they just knew which ones were their own. All 3 reported fable as the worst because it missed bugs that the others caught. Opus and astra essentially had the same number of severe bugs found, but the most bugs any one found was 13 while the 3 had 23 unique bugs found combined so it still pays to have multiple models review your code, at least at important milestones. They really do catch things that others miss. As for cost, astra and fable cost the same per token, but astra spent $6.67 while fable cost $7.48. Opus used up 50% of a 5hr limit on the $20 plan. Opus's overperformance might be a result of the Claude code harness, but fable and astra were both on opencode and fable did worse there while spending more. Also, those numbers are for both reviewing, and comparing the reviews, which included testing whether the findings were actual bugs or false positives. Astra was like $2.50 compared to $5.50 on Fable on just the reviewing step so it really outperformed fable at half the token usage. Opus also found one of its own bugs to be a false positive. Astra and fable found no false positives in any of the agents findings. Me personally, I will probably use astra when I need something smarter than opus, but my daily driver will still be some cheap Chinese model for implementation and opus for reviewing since I have the sub.

I have a $20 Claude sub and still pay per token on OC to do most of the work. I spend less on tokens than the Claude plan and I never hit a usage wall. I still use Claude opus for checking its work and writing good plans but just that usually consumes the 5hr limit.

The aspect ratio. It pisses me off that the android phones are just squares when open. This is actually a usable big screen.

Wait your brand new spark broke?

Just today and last night I tried using Deepseek flash on opencode and it kept getting stuck running a test, waiting for it to resolve, but the test would run tauri server and had no function to end the server after running the test. I had Claude check on the running processes every few minutes to free it.

Those exist already. what tf you gonna train with maybe 200 words