Hashan

Digital Marketing Professional
My tweets are purely my own opinions and do not represent any other entity.

[Hashan @hashanwick]

@MrAhmadAwais @0xSero 0.45B tokens usually would mean about $35-40 worth of tokens. Not $6. (Of course the actual cost is just a couple of dollars with that $1 offer). This $6 value just seems like a really inefficient use of tokens. Massively input heavy, hardly any outputs/real work product.

[Hashan @hashanwick]

@Aditya_181105 I wouldn't say either. DS V4 Pro is decent, GLM5.1 is very good too. Others are a step behind. Still, none of them production ready. They make too many mistakes, leaves garbage, leaves holes in the code, gets datafloes wrong, etc. AND hallucinates

[Hashan @hashanwick]

@Saanvi_dhillon And a forked and customised memory layer for Agents

[Hashan @hashanwick]

@Saanvi_dhillon Level 6:
Command Code with DeepSeek v4
Codex for GPT 5.5
Antigravity to make use of free Gemini quota
Claude Code with the GLM Coding Plan
Freebuff for free quota for CHANGELOG writing and documentation.
zFlow, superpowers and 9 arms Agent skills
A few MCPs and CLI tools

[Hashan @hashanwick]

@Gosleepriya @antoniomartinl ×3 simplest pattern to see

[Hashan @hashanwick]

@otokyo Scream and run like a girl! 🤣🤣

[Hashan @hashanwick]

@nalinrajput23 🤖👨‍💻

[Hashan @hashanwick]

@yadavji_codes Nope. I won't say it! There's one but I won't say what!

[Hashan @hashanwick]

@CodeWithAmann 64GB I'd say a decent amount. 128GB is what you need to comfortably run a decent MoE model locally while running the browsers, dev servers, docker and what not.

[Hashan @hashanwick]

@gregpr07 @itsmehatef Fair play! 😂

[Norveçli @norveclifinance]

DeepSeek just popped the American AI bubble.

Not by killing AI.

By killing the fantasy of unlimited AI pricing power.

DeepSeek V4 Pro:
Input: $0.435 per 1M tokens
Output: $0.87 per 1M tokens

OpenAI GPT-5.5:
Input: $5.00
Output: $30.00

Claude Opus 4.7:
Input: $5.00
Output: $25.00

Claude Sonnet 4.6:
Input: $3.00
Output: $15.00

DeepSeek is roughly:

11.5x cheaper than GPT-5.5 on input
34.5x cheaper than GPT-5.5 on output

28.7x cheaper than Claude Opus on output
17.2x cheaper than Claude Sonnet on output

If a model is “good enough” at 1/20th or 1/30th the cost, margins will compress faster than Wall Street expects.

AI is not dead.

But the AI bubble just lost its pricing power.

$NVDA $MU $AMD $SMCI $AVGO $ARM $MSFT $GOOGL $META $AMZN $ORCL $PLTR $CRWV $NBIS

#AI #DeepSeek #AIBubble #ArtificialIntelligence #Semiconductors #HBM #DRAM #Nvidia #Micron #AMD #TechStocks #Nasdaq #StockMarket #Investing #WallStreet #YapayZeka #Borsa

[Hashan @hashanwick]

@Im_IrushiK 1. Codex.
2. Gemini Pro
3. Cursor

I personally have Codex Pro, Gemini Pro, Command Code, GLM Coding Plan. I can get rough about 3B-3.5B tokens per month with this

[Hashan @hashanwick]

Permanently switching my harness to Command Code!

[Ahmad Awais @MrAhmadAwais]

@deepseek_ai LFG!! The $1 Go plan which stretches to $40 @CommandCodeAI for DSv4 Pro is now permanent. This is AI for everyone.

[Hashan @hashanwick]

@deepseek_ai @MrAhmadAwais will the $1 subscription 4x DeepSeek Pro offer extend beyond May?

[Hashan @hashanwick]

@almostme2696 @deepseek_ai What's the issue anyway. Its the same whether US has all my data or China has all my data. Neither would prioritise me and both would use me exactly the same way for their benefit. China at least cheaper...

[Hashan @hashanwick]

@Jesii_ca_M Too many people would lose regardless whatever the decision I make here so I won't give it to either of them. Equality

[Hashan @hashanwick]

@dev_maims ChatGPT right now. Used to be Claude. Grok is definitely a no go and Gemini is just too weird!

[Hashan @hashanwick]

@Samaytwt Aaaand now its back!

[Hashan @hashanwick]

@CleavageCrumbs @TaraBull Spot on

[Hashan @hashanwick]

@TaraBull Complete and utter nonsense. If you are running a restaurant, pay your staff.