diff --git a/cortex-js/src/infrastructure/commanders/models/model-start.command.ts b/cortex-js/src/infrastructure/commanders/models/model-start.command.ts
index 7c9268031..0e4aabae2 100644
--- a/cortex-js/src/infrastructure/commanders/models/model-start.command.ts
+++ b/cortex-js/src/infrastructure/commanders/models/model-start.command.ts
@@ -4,7 +4,7 @@ import { exit } from 'node:process';
 import { CortexUsecases } from '@/usecases/cortex/cortex.usecases';
 import { SetCommandContext } from '../decorators/CommandContext';
 import { ContextService } from '@/infrastructure/services/context/context.service';
-import { createReadStream, existsSync, statSync, watchFile } from 'node:fs';
+import { existsSync } from 'node:fs';
 import { FileManagerService } from '@/infrastructure/services/file-manager/file-manager.service';
 import { join } from 'node:path';
 import { Engines } from '../types/engine.interface';
@@ -14,11 +14,12 @@ import { isRemoteEngine } from '@/utils/normalize-model-id';
 import { downloadProgress } from '@/utils/download-progress';
 import { CortexClient } from '../services/cortex.client';
 import { DownloadType } from '@/domain/models/download.interface';
+import { printLastErrorLines } from '@/utils/logs';
 
 type ModelStartOptions = {
-  attach: boolean;
   preset?: string;
 };
+
 @SubCommand({
   name: 'start',
   description: 'Start a model by ID.',
@@ -77,16 +78,17 @@ export class ModelStartCommand extends BaseCommand {
       await downloadProgress(this.cortex, undefined, DownloadType.Engine);
     }
 
-    // Attached - stdout logs
-    if (options.attach) {
-      this.attachLogWatch();
-    }
-
     const parsedPreset = await this.fileService.getPreset(options.preset);
 
+    const startingSpinner = ora('Loading model...').start();
+
     await this.cortex.models
       .start(modelId, parsedPreset)
-      .then(() => options.attach && ora('Model is running...').start());
+      .then(() => startingSpinner.succeed('Model loaded'))
+      .catch(async (error) => {
+        startingSpinner.fail(error.message ?? error);
+        printLastErrorLines(await this.fileService.getLogPath());
+      });
   }
 
   modelInquiry = async () => {
@@ -104,16 +106,6 @@ export class ModelStartCommand extends BaseCommand {
     return model;
   };
 
-  @Option({
-    flags: '-a, --attach',
-    description: 'Attach to interactive chat session',
-    defaultValue: false,
-    name: 'attach',
-  })
-  parseAttach() {
-    return true;
-  }
-
   @Option({
     flags: '-p, --preset <preset>',
     description: 'Apply a chat preset to the chat session',
@@ -121,38 +113,4 @@ export class ModelStartCommand extends BaseCommand {
   parseTemplate(value: string) {
     return value;
   }
-
-  /**
-   * Attach to the log file and watch for changes
-   */
-  private async attachLogWatch() {
-    const logPath = await this.fileService.getLogPath();
-    const initialSize = statSync(logPath).size;
-    const logStream = createReadStream(logPath, {
-      start: initialSize,
-      encoding: 'utf-8',
-      autoClose: false,
-    });
-    logStream.on('data', (chunk) => {
-      console.log(chunk);
-    });
-    watchFile(logPath, (curr, prev) => {
-      // Check if the file size has increased
-      if (curr.size > prev.size) {
-        // Calculate the position to start reading from
-        const position = prev.size;
-
-        // Create a new read stream from the updated position
-        const updateStream = createReadStream(logPath, {
-          encoding: 'utf8',
-          start: position,
-        });
-
-        // Read the newly written content
-        updateStream.on('data', (chunk) => {
-          console.log(chunk);
-        });
-      }
-    });
-  }
 }
diff --git a/cortex-js/src/infrastructure/commanders/run.command.ts b/cortex-js/src/infrastructure/commanders/run.command.ts
index 086e6f701..1dcd88789 100644
--- a/cortex-js/src/infrastructure/commanders/run.command.ts
+++ b/cortex-js/src/infrastructure/commanders/run.command.ts
@@ -15,6 +15,7 @@ import { CortexClient } from './services/cortex.client';
 import { DownloadType } from '@/domain/models/download.interface';
 import { isLocalFile } from '@/utils/urls';
 import { parse } from 'node:path';
+import { printLastErrorLines } from '@/utils/logs';
 
 type RunOptions = {
   threadId?: string;
@@ -109,8 +110,10 @@ export class RunCommand extends BaseCommand {
         if (options.chat) this.chatClient.chat(modelId, options.threadId);
         else console.log("To start a chat session, use the '--chat' flag");
       })
-      .catch((e) => {
+      .catch(async (e) => {
         startingSpinner.fail(e.message ?? e);
+
+        printLastErrorLines(await this.fileService.getLogPath());
       });
   }
 
diff --git a/cortex-js/src/utils/huggingface.ts b/cortex-js/src/utils/huggingface.ts
index 27908f0f3..9df80eeeb 100644
--- a/cortex-js/src/utils/huggingface.ts
+++ b/cortex-js/src/utils/huggingface.ts
@@ -225,9 +225,9 @@ export async function getHFModelMetadata(
     const stopWord: string = metadata['tokenizer.ggml.tokens'][index] ?? '';
     const name = metadata['general.name'];
     const contextLength = metadata['llama.context_length'] ?? 4096;
-    const ngl = (metadata['llama.block_count'] ?? 32) + 1
+    const ngl = (metadata['llama.block_count'] ?? 32) + 1;
     const version: number = metadata['version'];
-    
+
     return {
       contextLength,
       ngl,