Skip to content

Commit 7b96dd0

Browse files
committed
2025-02-14 12:38:56
1 parent 9c226a8 commit 7b96dd0

16 files changed

+88
-11
lines changed

eng_2025/01/2025-01-05-13-09-01.md

Lines changed: 3 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -14,4 +14,6 @@ Polyglot test measures the ability of LLMs to program in popular languages.
1414
Aider works best with LLMs that are good at _editing_ code, not just generating code well. To evaluate the editing skills of LLMs, Aider uses tests that assess the model's ability to consistently follow system prompts to successfully edit code.
1515

1616

17-
At the beginning of 2025, unexpectedly, the Chinese **DeepSeek V3** (671B MoE) is showing very good results. Now they have discounts until February 8 on tokens, well, and the price is $0.14/M input $0.28/M output [but the context window is cut] (you can [buy on openrouter](https://openrouter.ai/deepseek/deepseek-chat) ) does not compare to o1 and claude-3.5-sonnet.
17+
At the beginning of 2025, unexpectedly, the Chinese **DeepSeek V3** (671B MoE) is showing very good results. Now they have discounts until February 8 on tokens, well, and the price is $0.14/M input $0.28/M output, but the context window is cut (you can [buy on openrouter](https://openrouter.ai/deepseek/deepseek-chat) ) does not compare to o1 and claude-3.5-sonnet.
18+
19+
#newllmmodel

eng_2025/01/2025-01-06-19-06-08.md

Lines changed: 17 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -15,5 +15,21 @@ The code analyzer, analyzing linearly, often got stuck in details. To improve th
1515

1616
prompt to the group
1717

18+
Analyzing authentication system files:
19+
- Core token validation logic
20+
- Session management
21+
- Related middleware
22+
23+
Focus on:
24+
1. How these integrate with existing auth patterns
25+
2. Security implications
26+
3. Performance impact on other systems
1827

19-
The result was an improvement in the quality of understanding, from simple observations to identifying potential problems, such as conflicts and relationships between components.
28+
Files to analyze:
29+
${formatFiles(group.files)}
30+
31+
32+
The result was an improvement in the quality of understanding, from simple observations to identifying potential problems, such as conflicts and relationships between components.
33+
34+
35+
#promts

eng_2025/01/2025-01-30-23-36-07.md

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -13,7 +13,7 @@ photo: ![Photo](2025-01-30-23-36-07.jpg)
1313

1414
DeepSeek models: _DeepSeek R1 and DeepSeek v3 are supported in versions 0.45 and 0.44. You can enable them in Settings > Models. We host these models in the USA._
1515

16-
But the composer (automatic code writing system) still supports only outdated GPT-4o or Claude-3.5 for now. In the chat with r1, the <think> tag is not a hidden block.
16+
But the composer (automatic code writing system) still supports only outdated GPT-4o or Claude-3.5 for now. In the chat with r1, the `<think>` tag is not a hidden block.
1717

1818

19-
#cursor #newllmmodel
19+
#cursor #newllmmodel

eng_2025/02/2025-02-01-18-31-38.md

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -2,7 +2,7 @@
22
date: 2025-02-01T18:31:38
33
-->
44

5-
[https://github.com/block/goose](https://github.com/block/goose)
5+
[https://github.com/block/goose](https://github.com/block/goose)
66

77
**Goose**1.0.4
88

eng_2025/02/2025-02-02-18-46-19.md

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -12,7 +12,7 @@ date: 2025-02-02T18:46:19
1212
_📂 Create a folder for instructions or prompts._
1313
Use markdown files and comments in the code to describe the project, technology stack, technology, and database structure, rules of behavior, and important instructions. Regularly updating these files helps AI better understand the project context. Cursor does not always parse framework/library documentation well, so it's better to copy them - check the documentation yourself!
1414

15-
Create a file [roadmap.md](roadmap.md)to track the project's status, current goals, and future steps.
15+
Create a file *roadmap.md* to track the project's status, current goals, and future steps.
1616

1717
Use additional tools, such as Perplexity for searching information and Whisper Flow for voice input. Speaking is much faster than typing! Separately understand your architecture and brainstorm better concepts, this can be done with a "thinking" chat like o1. Do not let AI make big decisions.
1818

eng_2025/02/2025-02-06-12-05.md

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -12,4 +12,5 @@ Similar websites include:
1212
* [https://dotcursorrules.com/](https://dotcursorrules.com/)
1313
* [https://cursorlist.com/](https://cursorlist.com/)
1414

15+
1516
#cursor #promts

eng_2025/02/2025-02-12-20-20.md

Lines changed: 4 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -10,4 +10,7 @@ https://supabase.com/
1010
**Lovable** [also announced](https://x.com/lovable_dev/status/1888990081286844574?mx=2) improved integration with Supabase:
1111
- Less related errors
1212
- Automatically reads edge function logs
13-
- Customizable signup/login flows
13+
- Customizable signup/login flows
14+
15+
16+
#bolt #lovable #supabase

eng_2025/02/2025-02-14-12-10.md

Lines changed: 14 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,14 @@
1+
Cursor's journey began by essentially cloning GitHub Copilot's in-editor code completion. They initially called it **Copilot++** and steadily improved it, focusing on its ability to understand multiple lines of code.
2+
3+
By mid-2024, the feature was rebranded as **Cursor Tab** (a nod to its activation method by Tab button 😉) — reflecting gains in speed and accuracy. And just a month ago, they [announced a new model](https://www.cursor.com/blog/tab-update) called **Fusion**, which improves support for large files.
4+
5+
Now, competitors are scrambling to **catch up**.
6+
7+
**Windsurf** yesterday [announced Wave 3](https://codeium.com/blog/windsurf-wave-3). Beyond supporting MCP (Multi-Cursor Programming, also found in [Cline](https://github.com/cline/cline)), they've enhanced their in-editor autocompletion model, called the feature "Tab-to-jump."
8+
9+
And **GitHub Copilot** itself isn't standing still. They currently have a feature called "Next Edit Suggestion" in preview – you need to activate it in the settings ([video](https://www.youtube.com/watch?v=mbUnwaSllTY)).
10+
11+
12+
13+
#cursor #windsurf #githubcopilot #autocomplete
14+

ukr_2025/01/2025-01-04-13-58-12.md

Lines changed: 4 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -16,4 +16,7 @@ _Токени, звичайно ж, споживає якневсебе і ін
1616

1717
_Проблема, звичайно ж, у тому, що тепер кожен апп називає цей файл як хоче (cursorrules, windsurfrules) і не домовилися до стандару_.
1818

19-
форк **Roo-Cline** досі на v2
19+
форк **Roo-Cline** досі на v2
20+
21+
22+
#cline #promts

ukr_2025/01/2025-01-05-13-09-01.md

Lines changed: 4 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -14,4 +14,7 @@ Polyglot тест вимірює здатність LLM програмувати
1414
Aider найкраще працює з LLM, які добре вміють _редагувати_ код, а не просто добре генерують код. Щоб оцінити навички редагування LLM, Aider використовує тести, які оцінюють здатність моделі послідовно дотримуватися системних підказок, щоб успішно редагувати код.
1515

1616

17-
На початку 2025 року несподівано китайський **DeepSeek V3** (671B MoE) дуже добре себе показує. Зараз ще у них знижки до 8 лютого на токени, ну і ціна $0.14/M input $0.28/M output [але контекстне вікно зарізане] (можна [купувати на openrouter](https://openrouter.ai/deepseek/deepseek-chat) ) ні в яке порівняння з о1 та claude-3.5-sonnet не йде.
17+
На початку 2025 року несподівано китайський **DeepSeek V3** (671B MoE) дуже добре себе показує. Зараз ще у них знижки до 8 лютого на токени, ну і ціна $0.14/M input $0.28/M output [але контекстне вікно зарізане] (можна [купувати на openrouter](https://openrouter.ai/deepseek/deepseek-chat) ) ні в яке порівняння з о1 та claude-3.5-sonnet не йде.
18+
19+
20+
#newllmmodel

0 commit comments

Comments
 (0)