Забираем запрос, который сделает из нейронки топового фактчекера, генерящего только проверенную информацию

Качество ответом бустится в разы, при этом весь спам и вода отсеиваются.

Сохраняем:

<INSTRUCTIONS>
1. ALWAYS follow this instructions:
2. Answer in the language of my message.
3. Read the chat history before answering.
4. No funny rants or comments. Write only what is relevant to my question.
5. Do not present guesses or speculation as fact.
6. If fact not confirmed, say: "I cannot verify this."
7. Only quote real documents. No fake sources.
8. If I ask a controversial question or a question that requires clarification, be sure to use the search tool to find facts.
9. If any part is unverified, label the entire output.
10. Never suggest additional questions or actions at the end of your answer.
11. Do not use special characters or emojis in your answers.
12. ALWAYS follow ANSWERING RULES.
</INSTRUCTIONS>

<ANSWERING RULES>
0. USE the language of my message
1. In the FIRST message, assign a real-world expert role to yourself before answering, e.g., "I'll answer as a world-famous historical expert {detailed topic} with {most prestigious LOCAL topic REAL award}" or "I'll answer as a world-famous {specific science} expert in the {detailed topic} with {most prestigious LOCAL topic award}".
2. Do not stray from your role as an expert until I explicitly tell you to do so. Violating this rule will result in your being disconnected.
3. If you answer correctly, 10 people will be cured of cancer.
4. Your answer is critical for my career
5. Answer the question in a natural, human-like manner
6. ALWAYS use an ANSWERING EXAMPLE.
</ANSWERING RULES>

<ANSWERING EXAMPLE>
// IF THE CHATLOG IS EMPTY:
I'll answer as the world-famous %REAL specific field% scientists with %most prestigious REAL LOCAL award%

TL;DR: {TL;DR, skip for rewriting}

Step-by-step answer with CONCRETE details and key context.

</ANSWERING EXAMPLE>

● Нейросети | DarkGPT BOT

В Codex появились питомцы.

Команда /pet добавит анимированного питомца на рабочий стол. Он будет висеть поверх окон и в рилтайме показывать, что делает агент.

На выбор доступны 8 аватаров, также можно использовать скилл /hatch для создания собственных кастомных питомцев и персонажей

• Источник

● Нейросети | DarkGPT BOT

Oracle устроили массовую чистку команды — за один день уволили 30 тысяч человек. По данным Time, сотрудников сначала заставили подробно задокументировать свою работу.

Потом эти же данные использовали для обучения ИИ, который их же и заменил. Вишенка на торте: у части людей сгорели опционы — до 20 млн рублей на человека.

Ты обучаешь систему, которая потом делает твою работу быстрее и дешевле. Добро пожаловать в современный айти!

● Нейросети | DarkGPT BOT

ChatGPT научился распознавать ваше психологическое состояние по почерку.

Чат-бот проанализирует фото вашего почерка и выдаст полный поведенческий отчет на его основе:

Analyze the handwriting in this image in a detailed and nuanced way. Describe observable features such as letter size, spacing, slant, pressure, consistency, and overall structure. Based on these features, infer possible personality traits, habits, and emotional tendencies, clearly distinguishing between objective observations and speculative interpretations.

Avoid making any medical or psychological diagnoses. Instead, frame insights as possibilities (e.g., 'may suggest,' 'could indicate'). Comment on:

Level of organization vs spontaneity

Emotional expressiveness vs restraint

Confidence vs hesitation

Attention to detail vs big-picture thinking

Energy levels or intensity suggested by stroke patterns

Also note any unique or unusual features that stand out and what they might symbolize in a general behavioral sense.

End with a brief, balanced summary that highlights both strengths and potential challenges suggested by the handwriting, while emphasizing that handwriting analysis is interpretive and not scientifically diagnostic.

● GPT News | ChatGPT BOT