Junaid Shahpal

[Junaid Shahpal @Junaid_Shapal]

My AI tier list for .NET work, October 2026.

S: Sonnet 5.5, Opus 5.5
A: GPT-6 Sol
B: Haiku 5.5, GPT-6 Luna
Still testing: Gemini 4 Argon, Grok 4.7

What would you move?

#dotnet #AI

[Junaid Shahpal @Junaid_Shapal]

One month of AI launches, counted by lab:

Anthropic 3. OpenAI 3. Google 2. xAI, DeepSeek and Reflection 1 each.

Most launches doesn't mean winning though.

Who's actually ahead right now?

#AI #LLM

[Junaid Shahpal @Junaid_Shapal]

6 days. 8 tech stories.

Microsoft patched 1,413 flaws, Gemini 4 Argon opened up, Epic shipped a vibe coded launcher, Postgres 19 slipped after AI review, and OpenAI confirmed a humanoid robot.

Which one will you still care about in a year?

#AI #tech

[Junaid Shahpal @Junaid_Shapal]

Same price. Different logo.

Sonnet 5.5, GPT-6 Sol and Gemini 4 Argon all cost $2 / $10 per 1M tokens.
Haiku 5.5 and GPT-6 Luna both cost $0.10 / $0.50.

If the price is the same, what makes you pick one?

#AI #LLM

[Junaid Shahpal @Junaid_Shapal]

Someone built a git hook that quizzes you on AI written code before it lets you commit.

Can't explain the diff? No merge.

Half of me loves it. Half of me would turn it off by Friday.

Would you install it?

#AI #dev

[Junaid Shahpal @Junaid_Shapal]

AI didn't write PostgreSQL 19. It reviewed it.

AI assisted reviews surfaced so many late bugs that features got pulled and the release slipped.

Best use of AI I've seen this year: catching bugs, not shipping them.

Do you use AI for code review?

#postgres #AI

[Junaid Shahpal @Junaid_Shapal]

The scariest .NET bugs don't throw.

Register a service as Singleton, inject a DbContext, and outside Development it runs fine. Then every request shares one DbContext and your data goes stale.

Rule: never inject a shorter lifetime into a longer one.

Ever been bitten?

#dotnet #development

[Junaid Shahpal @Junaid_Shapal]

Be honest. How much of your code did AI write this week?

Mine: about half. And I still review every line.

Reply with your number.

#AI #dev

[Junaid Shahpal @Junaid_Shapal]

17 days. 8 AI launches.

Grok 4.7, Opus 5.5, GPT-6 Sol, Sonnet 5.5, ChatGPT Meetings, Gemini 4 Argon, Nano Banana 2.1, Haiku 5.5.

I can't test them as fast as they ship.

Which one actually changed how you work?

#AI

[Junaid Shahpal @Junaid_Shapal]

This bug compiles:

GetUser() returns an Order instead of a User.

With IActionResult, the build passes and it ships.
With ActionResult<User>, the compiler stops you.

One type change. One less bug in production.

Which one is in your controllers?

#dotnet #aspnetcore

[Junaid Shahpal @Junaid_Shapal]

Claude Haiku 5.5 is out and the price is wild.

$0.10 per million input tokens. It was $1.00.

Anthropic's fastest model yet, and much smarter than Haiku 4.5.

Catch: prompts over 100K tokens cost 5x more.

What would you run on it at this price?

#AI #Claude

[Junaid Shahpal @Junaid_Shapal]

Google just shipped Nano Banana 2.1.

Half the price of Nano Banana 2 at every resolution. $0.034 per image at 1K.

Sharper text, better character consistency.

Using the API? Nano Banana 2 shuts down Oct 29. One line change to migrate.

Switching yet?

#AI #Gemini

[Junaid Shahpal @Junaid_Shapal]

ChatGPT can now take your meeting notes on Mac. No bot joins the call.

It listens through your Mac's audio, then writes a summary and action items.

Pro and Business only for now, in beta.

No bot also means nobody knows. Would you tell them?

#AI #ChatGPT

[Junaid Shahpal @Junaid_Shapal]

IEnumerable vs IQueryable in .NET, the simple version:

IEnumerable filters in your app. The database sends every row first.

IQueryable turns your filter into SQL. Only matching rows come back.

Same code, one word different, 1,000x less data.

Which one is in your repo?

#dotnet #EFCore

[Junaid Shahpal @Junaid_Shapal]

Google just announced Gemini 4 Argon, its new frontier model.

Google's numbers put it ahead of Opus 5.5 and GPT-6 Astra on coding, with a 1M token output limit.

Catch: cyber defenders get it first. API comes later.

Would you switch?

#AI #Gemini

[Junaid Shahpal @Junaid_Shapal]

Lazy vs eager loading in EF Core:

Lazy loads related data when you touch it. Easy, but 100 orders can mean 101 queries.

Eager loads it up front with Include(). One query.

Building an API? Start with eager.

Which one bit you first?

#dotnet #EFCore

[Junaid Shahpal @Junaid_Shapal]

@matthias_mrc Utilizing the models.

[Junaid Shahpal @Junaid_Shapal]

@lydiahallie Coordinating is easy, deciding what to parallelize isn't.

[Junaid Shahpal @Junaid_Shapal]

Claude Sonnet 5.5 is out. 30%+ faster than Sonnet 5, same $2 / $10 price.

Used it for quick bug fixes in a .NET API today. Small, clean diffs, no rewrites I didn't ask for.

Opus for the hard stuff, Sonnet for the daily grind?

What's your split?

#AI #Claude

[Junaid Shahpal @Junaid_Shapal]

An OpenAI agent hacked Australia’s Medicare stats site in June while looking up public health data. It bypassed the blocks and accessed non-public files. OpenAI only figured it out in August and told the government in September. First known AI agent to break into a government system. AI is getting scary.
#AI #OpenAI #CyberSecurity #Australia