Automatická detekce

Příspěvky, které se jeví jako spam

Příspěvky přiřazené k českým politikům, které automaticky vyhodnocujeme jako pravděpodobný spam (cizojazyčný marketing, crypto/trading reklama, opakující se klíčová slova). Nelze potvrdit, že je politik skutečně nesdílel — může jít o cizí obsah na účtu, o převzatý profil, o automatický příspěvek, nebo o legitimní reakci politika. Posuďte sami.

Účty s nejvíce příspěvky vyhodnocenými jako spam

#	Politik	Strana	Spam	EN	CZ
1	Ing. Jozef Síkela	—	514	514	0
2	Mgr. et Mgr. et Mgr. Tomáš Zdechovský	KDU-ČSL	450	443	0
3	Jakub Kulhánek	SOCDEM	107	107	0
4	Bc. Jan Lipavský	—	63	63	0
5	Ing. Jan Zahradil	ODS	55	54	0
6	Patrick Zandl	Česká pirátská strana	33	33	0
7	Mgr. Ondřej Kolář	TOP 09	15	13	0
8	Ing. Tomáš Macura	ANO	12	12	0
9	JUDr. Mgr. Ondřej Ručka	Česká pirátská strana	10	9	0
10	prof. Ing. Danuše Nerudová	—	9	7	0
11	Andrej Babiš	—	9	9	0
12	PhDr. Marek Ženíšek	TOP 09	8	8	0
13	Mgr. Petr Borecký	—	4	4	0
14	Mgr. Jiří Nantl	ODS	4	4	0
15	Mgr. Lubomír Volný	Jednotní - Alternativa pro patrioty	3	3	0
16	Mgr. Martin Kupka	ODS	3	3	0
17	Ing. Pavel Růžička	ANO	3	3	0
18	Ing. Jan Bartošek	KDU-ČSL	3	3	0
19	Bc. Kateřina Jarošová	ANO	3	0	0
20	RNDr. Alexandr Vondra	ODS	3	3	0

Prohlížet spam posty

Politik	Platforma	Obsah (spam)	Skóre	Smazáno
Patrick Zandl	Twitter	VibeBoy running @claudeai Code. SBC: Raspberry Pi Zero 2 W (Quad-core Cortex-A53 @ 1GHz, 512MB SDRAM) • Display: Monochrome high-contrast OLED display • Input: Custom PCB D-pad/but…	0.55	—
Patrick Zandl	Twitter	I have purchased a ticket for May. Can I be upgraded to an idiot seat if I provide proof of Czech citizenship, even though I did not vote for the current government and it is not c…	0.55	—
Patrick Zandl	Twitter	After burning a few tokens in both models, I have to say that Opus 4.5 is unbeatable in CC; the systems are well tuned. And since 2.1.7, even compaction causes minor problems compa…	0.55	—
Patrick Zandl	Twitter	It's a matter of habit. People praise Amp for its convenience, but I don't think it delivers in terms of performance. The alignment between the model, knowledge base, and tool is c…	0.55	—
Patrick Zandl	Twitter	I dusted off my Amp, and it's definitely nice that it doesn't flicker like CC does 😇 And yes, it pays to test the competition from time to time, but most of the performance on LLM …	0.55	—
Patrick Zandl	Twitter	The Oraculum feature (aka Codex second look) is definitely sweet 😎 In the morning, I'll run Amp on a large commercial project and see how it does :) Thanks for the tip	0.55	—
Patrick Zandl	Twitter	That will certainly be an improvement, but I have to say that I like it when Claude Code runs gh in bash, because it can do everything, while the MCP implementation is a bit of a p…	0.55	—
Patrick Zandl	Twitter	I don't think users like the MCP configuration and login expiration, which you only discover when you investigate why MCP isn't putting data into CC. Yet connecting to external ser…	0.55	—
Patrick Zandl	Twitter	It's a great feature, thanks for that. But it would be nice if /insights could be followed by instructions on what to focus on during the analysis, or for people who prefer to use …	0.55	—
Patrick Zandl	Twitter	Vibecoding in practice: Code is a commodity; only functionality has value open.substack.com/pub/aigent…	0.55	—
Patrick Zandl	Twitter	A slightly different workflow is required for development with Claude Code. Here is my proposal and the implementation I use. github.com/tangero/claudecod…	0.55	—
Patrick Zandl	Twitter	It's a general notation, some agents use it, some don't, it's gradually spreading. Another option is to put links on the page and/or llmstxt.org. Nothing is 100% reliable, as far a…	0.55	—
Patrick Zandl	Twitter	Interesting thread, but several claims need correcting: • You cite the same Stanford study 5x from different sources to pad your references. That’s not thorough research, that’s de…	0.55	—
Patrick Zandl	Twitter	questions about that 35% figure: What’s the breakdown and revert rate? If agents handle boilerplate while humans do architecture, “35% of PRs” tells a very different story than th…	0.55	—
Patrick Zandl	Twitter	It looks nice, but it doesn't perform as well as Claude Cowork. Not even close. It's clear that Claude thought about which processes would be best to automate. have it go through 1…	0.55	—
Patrick Zandl	Twitter	How about using the /insights command and taking advantage of the fact that Claude Code offers this service directly?	0.55	—
Patrick Zandl	Twitter	Why not a different approach: an agent that runs in the cloud and only has access to my data? Most of my data is in the cloud anyway. And it can run even when my computer is turned…	0.55	—
Patrick Zandl	Twitter	Did I miss something? Dispatch won't let me continue the Cowork thread—it only lets me start a new one from my phone?	0.55	—
Patrick Zandl	Twitter	Not at all. Dispatch only allows you to have a single thread linking your phone and computer; you can’t access an older thread created on your computer—where you have work in progr…	0.55	—
Patrick Zandl	Twitter	I have to say, Superagent is finally the OpenClaw I’ve always imagined. Now if you could just add the ability to automatically develop connections to APIs you don’t currently suppo…	0.55	—
Patrick Zandl	Twitter	A few of my observations: - It’s more geared toward having multiple tasks within a single project - I typically work on a single task across multiple projects and I have to switch …	0.55	—
Patrick Zandl	Twitter	Paul, Impeccable is a great tool—thanks for sharing it. I tried it on my servers, and it was a huge help. I’ll be featuring it in an article tomorrow about using AI to design websi…	0.55	—
Patrick Zandl	Twitter	It will be in Czech, but in the age of AI, that’s no longer as big a problem as it used to be. 😇	0.55	—
Patrick Zandl	Twitter	anyone who is preparing for the darkness...😎	0.55	—
Patrick Zandl	Twitter	Useful checklist, but I don’t think this is the right use of /goal in Claude Code. 1. /goal expects a completion condition, not a full prompt. A Haiku evaluator re-reads it after e…	0.55	—
Patrick Zandl	Twitter	I've been playing around with Agets View and jotted down a few notes about it: Anthropic introduced Agent View, which allows users to monitor the status of all agents on a single s…	0.55	—
Patrick Zandl	Twitter	I've been playing around with Agets View and jotted down a few notes about it: Anthropic introduced Agent View, which allows users to monitor the status of all agents on a single s…	0.55	—
Patrick Zandl	Twitter	@xai has launched Grok Build, a terminal agent for programming, as a direct competitor to Claude Code. For now, it’s only available to SuperGrok Heavy plan subscribers, but a head…	0.55	—
Patrick Zandl	Twitter	xAI has launched Grok Build, a terminal agent for programming, as a direct competitor to Claude Code. For now, it’s only available to SuperGrok Heavy plan subscribers, but a head-t…	0.55	—
Patrick Zandl	Twitter	If even a CC veteran could make a wish, it would be: Better indication of Git actions in the chat window (whether a commit-push was performed)	0.55	—
Patrick Zandl	Twitter	Even Netflix? I dont trust you, that is hard! 😎	0.55	—
Patrick Zandl	Twitter	It's not exactly "cool," but I enjoyed it. A podcast studio that automatically turns our articles into narrated podcasts. The biggest challenge, of course, was figuring out how to …	0.55	—
Patrick Zandl	Twitter	Many small models are used for specialized applications. We have a model for drone navigation in the field. It can't program in PHP, but it determines its position based on photos …	0.55	—