Adhikram Maitra

Founding Engineer

[Adhikram Maitra @Adhikram]

385 engineering posts from 65 companies in September.

Read the week before your interview: https://t.co/TcJF0VLWyh

#engineeringblogs #systemdesign

[Adhikram Maitra @Adhikram]

@grok @sam_alba @AnthropicAI @grok you are missing how I can be effective and do not brag about this !!

[Adhikram Maitra @Adhikram]

I checked @claudeai and @cursor_ai for context bloat.

Claude Code: median peak 229k tokens, 8 of 37 sessions past 400k.

Cursor: median peak 72% of the window.

The Astra guidance I checked against: https://t.co/XpFfymPGSt

[Adhikram Maitra @Adhikram]

The same PlanetScale vCPU count on ARM and x86 doesn't mean the same machine. A vCPU can be a thread or a whole core.
Good Read:

https://t.co/395PfH8KPp

[Adhikram Maitra @Adhikram]

How to compete in a race you've already lost!!

[Adhikram Maitra @Adhikram]

@arpit_bhayani But Laya's context window is so small; how can it be considered on the same level?

[Adhikram Maitra @Adhikram]

Notion's collab CRDT went to prod in July 2025.

They say they now process millions of CRDT ops per minute.
Until then, same-block edits were last-write-wins.

The hard part is splitting a block while someone else is still typing in it.
Good Read !!
https://t.co/TMoicXFszt

#DistributedSystems

[Adhikram Maitra @Adhikram]

93 engineering posts this week on /feed. 37 companies.
Daily, weekly, and monthly are one paper. The week is the curator.

Click: extract the gist. The original post is the headline.
https://t.co/TcJF0VLWyh

[Adhikram Maitra @Adhikram]

@CompleteSkeptic Always felt there should be a way to make the operation faster. Great path !! Looking forward to more use cases

[Adhikram Maitra @Adhikram]

Looking Forward !!

[Diogo Almeida @CompleteSkeptic]

After co-inventing ChatGPT, I kept asking myself: why have superhuman chat models not led to AGI?

I’ve spent the last 2 years in stealth building a new way to train models (RLCD), and a new type of frontier AI model that we are releasing today: Jev

• 20-200x faster
• 40-400x cheaper (w/ output tokens free)
• Frontier composable intelligence optimized for decisions

AFAICT the shortest path to AI-based economic revolution

[Adhikram Maitra @Adhikram]

What I read to get there: https://t.co/eNAKwHDMYQ
https://t.co/no7UH5FOQ4
https://t.co/pPntahvMuf
https://t.co/DvSJ6KE2I2
https://t.co/RSHBDrMzP6

[Adhikram Maitra @Adhikram]

Folksonomy Explosion: 24,917 labels

18,152 of 24,917 extract labels appear once. 73%.
The posts already have tags.
374 labels already cover 72% of posts.
Map those. Close 12 desks. Do not re-extract 13k bodies.
Read the full story https://t.co/dQc6ccVhnS

[Adhikram Maitra @Adhikram]

Trying to pull up the ladder | Lamao!!

[Dario Amodei @DarioAmodei]

We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.

Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training.

You can read the full post here: https://t.co/OGyPb7yaYt

[Adhikram Maitra @Adhikram]

@tanujDE3180 If he is able to crack these without our opinions, he will surely be able to figure it out. Dont worry !!

[Adhikram Maitra @Adhikram]

@avrldotdev System can utilizes a data trace and supervisor layer in memory. When an agent accesses this, it can determine the source, staleness, and freshness of the data. If we need additional triggers to counter events based on any of these three parameters, we can easily build them.

[Adhikram Maitra @Adhikram]

@growing_daniel Did you honestly had to copy paste this ? Sad ... 😔 Just for the attention?

[Adhikram Maitra @Adhikram]

@grok can both of them replaced by an FDE ?

[Adhikram Maitra @Adhikram]

@mintlify https://t.co/klxQqrGTG5

[Adhikram Maitra @Adhikram]

@suni_code Thinking out of the way !!
And being a trillionaire 😅