IS AI PROGRESS SLOWING DOWN?

Hello Reader,
If you’ve experimented with AI, you might have hit the wall on its capabilities pretty early. Sometimes, AI just feels dumb and latest AI models don’t feel like they are doing anything better than the previous ones.

Last week I’ve stumbled upon a great video that explains the apparent slowdown, but more importantly dives into how will our experience with AI improve:

1. Specialization is the Future​
   OpenAI might not make the best models anymore, but they’re pouring resources into user-friendly products like ChatGPT. Anthropic, meanwhile, excels at API-driven models that are slightly better but lack polished consumer interfaces. The takeaway? The battle isn’t just about smarter AI; it’s about how effectively you can use it.
2. Faster and Cheaper Hardware​
   New chip companies are building specialized hardware that makes AI faster and cheaper. Their goal is less about making AI smarter and more about making it accessible to more people for more use cases.
3. Smarter Use of What We Have​
   Remember how late-gen games on the Xbox 360 looked way better than early ones, even though the console was the same? That’s what’s happening with AI. Companies are figuring out how to squeeze every ounce of potential from existing models, making them more useful and reliable.

What does it mean for testing?

We keep hearing about AI replacing testing, but I don’t see this happening anytime soon. However a tester being able to work effectively with AI tools might have the edge. It seems like we are all about to learn how to use this tool effectively and those who will get better, will do better.

What do you think about these trends? Are you optimistic, skeptical, or somewhere in between? Hit reply and let me know—I’d love to hear your thoughts!

Blogposts, discussions, events

Lies and trends in AI Is AGI right around the corner? Not even close. But it is evolving. A great video explaining what’s happening, from Steve Sewell ​Watch video →​

Vitest If you ever used Jest, then you should definitely check out Vitest. I’ve been recently able to play with it and I was pleasantly surprised with its capabilities. Read more →

Writing tests that fail My friend Artem wrote this refreshing take on failing tests. Failing test is good news! That is given that it gives you the information you need. ​Read more →​

Test automation tip

If cookie messages get in your way when testing, you can use the same trick as when logging in. Save the application state so that your application "remembers" your settings. Both Playwright and Cypress have these capabilities.

 1  // playwright
 2  import { test as setup, expect } from '@playwright/test';
 3  
 4  setup("user authentication", async ({ page }) => {  
 5    await page.goto('/');
 6    await expect(page.getByTestId('cookie-consent-message')).toBeVisible();
 7    
 8    await page.getByRole('button', { name: 'Accept' }).click();
 9    await expect(page.getByText('Cookie Preference Saved')).toBeVisible();
10    
11    await page.context().storageState({ path: 'playwright/.auth.json' });
12  });

 1  // cypress
 2  beforeEach(() => {
 3    cy.session('allow cookies', () => {
 4      cy.visit('/');
 5      cy.get('[data-cy=cookie-consent-message]')
 6        .should('be.visible')
 7      cy.contains('Accept').click();
 8      cy.contains('Cookie Preference Saved');
 9    });
10  });
11  
12  it('should not show cookie consent message anymore', () => {
13    cy.visit('/');
14    cy.get('[data-cy=cookie-consent-message]')
15      .should('not.exist')
16  });

Meme of the week

​

Filip Hric Teaching testers about development, and developers about testing

filip@filiphric.sk, Senec, Slovakia 90301
​Unsubscribe · Preferences​