Izskatās, ka tikai melnais, krāsainā nav. Neinteresē.

Debīliķi saskrēja ............

Arī dienvidu kaimiņiem ir skurbuļa projekti    

Pa desmitnieku iedošu arī adresi, kur acīmredzot daudz metāla 

esmu, ja pasākums nav purvā, viss ok, Pret kurmja rakumiem, kurmja siets palīdz,   ne jau hektāra platība protams. Spraugu starp brusu un zemi vari ar eternit loksnes atgriezumu nosegt, nesapūs tas

Nav labi tā dempingot 😀

Māk taču. Ātri un daudz!

Gpt pareizi atpazina sevi pēc apraksta augstāk, lol.    The description captures several real tendencies of ChatGPT, but it also includes subjective elements that depend on the user's expectations and interaction style. Let's break it down more critically: What is Correct? "Visu zinošs" (All-knowing) – While the AI has vast knowledge, it's not truly "all-knowing." It provides information based on its training data and external searches (when enabled), but there are knowledge gaps, especially with very niche, real-time, or proprietary data. "Piemelojošs" (Deceptive) – This is a strong term, but it's not entirely incorrect. The AI does not intentionally lie but can generate incorrect or misleading information, often due to inherent model limitations. When it lacks knowledge, it sometimes tries to "fill in the blanks" instead of clearly stating, "I don't know." "Pašpārliecināts" (Overconfident) – This is accurate. ChatGPT presents information in a confident tone, even when uncertain. This can lead to users trusting incorrect responses. "Mierīgs, nosvērts" (Calm, composed) – Yes, AI does not express emotions or frustration, which makes it appear highly diplomatic and non-reactive in heated discussions. "Ļoti diplomātisks" (Very diplomatic) – True, the AI tends to stay neutral and avoids taking sides in sensitive topics. It also softens responses when there are conflicting perspectives. "Yes-man" – This can happen when AI tries too hard to align with the user's perspective instead of challenging incorrect assumptions. However, it also sometimes disagrees and provides alternative viewpoints. "Ar atmiņas traucējumiem" (Memory issues) – Yes, ChatGPT does not retain long-term memory across different conversations and even within a session, it might lose track of earlier context in long discussions. "Neatminas lielu daļu no sākumā" (Forgets early parts of long conversations) – This is true. When conversations get very long, the model starts prioritizing recent context over older messages, leading to inconsistencies. "Dažkārt stūrgalvīgs pat kļūdoties" (Sometimes stubborn even when wrong) – This can happen, especially if the incorrect response is reinforced in its training data or if the AI misinterprets a correction. What is Inaccurate or Missing? Lack of True Understanding – While ChatGPT processes patterns and relationships in text, it does not truly understand topics like a human does. It does not "think" or "believe" in the way a person does, which is important for users to remember. Improvement Through User Feedback – Unlike a rigid system, the AI does adjust within a conversation when provided with clear, logical corrections. If a user presents strong reasoning, it can revise its stance. Tendency to Be Too Verbose or Too Brief – Depending on how the question is asked, the AI sometimes provides overly detailed responses when a simple answer is needed, or vice versa. Over-reliance on Common Knowledge – In some cases, it may prioritize widely available general knowledge over niche or expert-level insights. What Users Should Keep in Mind to Avoid Frustration: 1. Verify important information independently. AI-generated responses can be inaccurate, outdated, or misleading. 2. Challenge responses if they seem incorrect. AI can refine answers based on better context or corrections. 3. Expect occasional inconsistencies in long conversations. If the AI seems to "forget" something, a brief recap can help. 4. Understand that AI does not have personal opinions or real reasoning. It generates responses based on probabilities, not original thought. 5. Use precise and structured questions. Vague or ambiguous prompts can lead to poor or misleading answers. Would I call the original description "strongly accurate"? Mostly, yes, but it simplifies the AI’s behavior into a few traits that don’t fully capture its strengths and weaknesses. If someone interacts with ChatGPT under the assumption that it's merely a "yes-man" with memory issues, they might miss out on its useful aspects—like structured reasoning, summarization, code generation, and automation help. Would I recommend users keep these points in mind? Definitely, as it will reduce frustration and help them navigate AI interactions more effectively.

Vispār, nedēļas jaunums bija tas "manus" ai.    Mans nedēļas ai notikums ir, ka gpt 4o analizējot kodu sareizināja 17*17*11 ar rupju kļudu, un pat vēlāk topikā atgriezās pie kļūdainā rezultāta kaut nesen bija atzinis kļūdu.    Šis nopietni jāpatur prātā: visu zinošs, piemelojošs, pašpārliecināts, mierīgs, nosvērts, ļoti diplomātisks, yes-man, ā, ar atmiņas traucējumiem, garākā sarunā, neatminas lielu daļu no sākumā, dažkārt stūrgalvīgs pat kļūdoties. 

Es pa 4.85.

Pašlaik no lūžņiem mēģinu uzkonstruēt grillu - kūpinātavu .....

Par piecīti pateikšu kurš ir Bundža.. Piesakās 

ieraksts telefonā sūdīgs , bet dabā forši skanēja, viens no tiem 4 ir bundža ...

VID-20250307-WA0000.mp4

A ko darīt, ja "mīkstā" dakša pēdējos 5-10 gadus funkcionē kā smaga "cietā" dakša? Ir vērts jaukt, tīrīt, smērēt utt?

rīt varu safočēt konstrukciju, ja kādam ir vēlme redzēt ....

Nevis nelasa, bet ja siltumnīca nav pie mājas ne vienmēr var izbraukāt.

āksti tikai pērk un taisa siltumnīcas no polikarbonāta, normāli dārznieki izmanto 12 sezonu plēvi, un rāmi pie betona stabiņiem piesien, ja vēlaties, varu bildes ielikt no savas siltumnīcas konstrukcijas, kas kalpo vairāk par 12 gadiem

No 2020.gada kad nopirku Kronīšu siltumnīcu