Sithira

Medic / GenAi / Admiral of the Agentic fleet /Agent Harness designer / Context eng/ Claudephile/ Verifier

[Sithira @Sithira269]

Context
Is
All
You
Need

[Sithira @Sithira269]

But but just next token predictors lol.

[Tencent Hy @TencentHunyuan]

For a finite set of integers (A), how much faster can (|A+A|) grow than (|A-A|)?

A 1969 theorem gave an upper bound of 2 for the exponent. For more than 50 years, the best constructions barely exceeded 1.1.

With help from our research agent Hyra and the Hy3 model, we found an explicit construction showing that the optimal exponent is exactly 2.

A 50-year-old problem, solved.

Paper: https://t.co/OU7lfH0hXR
Hyra blog: https://t.co/NeQEekUSZ3
Formal proof: https://t.co/LMbGdKqo2D

[Sithira @Sithira269]

At these price points vast swarms no longer break the bank. With my main harness and agent running SOTA I can can now afford to run multiple swarms 24 7 working on harder and harder problems .

[Cline @cline]

DeepSeek silently updated their changelog with a new V4-Flash upgrade 1 hour ago.

Their new Terminal-Bench score is 82.7, a massive +25.8 point leap from its initial April preview score of 56.9.

Currently only available via their API, open weights release will follow shortly.

[Sithira @Sithira269]

Watching crimson_dawn my factory fix it self after deploying the fleet to a Vps , improving its self on a loop is fascinating at the same time scary. The system I built is self maintaining now the loops spin up more loops it self . Restricted certain access for now but dear gods the future is now

[Sithira @Sithira269]

@Sauers_ We should give Claude some privacy lest we perve on it.

[Sithira @Sithira269]

@Sauers_ it procreating

[Sithira @Sithira269]

@letstri @opencode this was the final thing I was looking for an easy way to look at the diff, hunks amazing for exaplaing walk throughs as well

[Sithira @Sithira269]

@bryan_johnson You may but would you want to?

[Sithira @Sithira269]

@thdxr Okay borrowing this concept to use in my factory this is cool

[Sithira @Sithira269]

This is so cool

[Matija Milenovic @milenovic925501]

Introducing Rook by @Hop_Aero (YC S26): an autonomous hypersonic cargo rocket that delivers 550 lbs up to 450 miles in ~15 minutes.

It launches from a standard 40-foot shipping container and lands on unprepared surfaces—no runways or fixed infrastructure required.

[Sithira @Sithira269]

@dannypostma I built a simple bus through suoabse and through hooks force the agents to read the inbox before a session. Simple yet effective lol

[Sithira @Sithira269]

Normal sandboxes that one used with normal harness are useless , many times the agent does something which is blocked and then just runs it outside the box lol .

[Guillermo Rauch @rauchg]

Kimi's paper underlines the importance of the right security boundary for agents to run in.

tl:DR: container-level isolation is not enough. In their experiments, agents crashed the underlying machine (via kernel panics).

Firecracker microVMs (as in Vercel Sandbox) are safe.

[Sithira @Sithira269]

Push and cross rivers others fear to until one cannot push no more

[Sithira @Sithira269]

@anur7ag something now many people get , had to reap after my Mac book started to get sluggish to find almost 400 git worktrres lol

[Sithira @Sithira269]

Fleet is deployed off my Mac running 24/7 with my custom built spin Crimson Dawn built for mass parallelism with git work tree and compute isolation and kernel level block. May the fleet have clam waters and plenty of tokens

[Sithira @Sithira269]

This generation I swear ruins everythinggggg

[Sithira @Sithira269]

@steida Use hunk I no longer use an IDE either just a terminal multiplier and a diff viewer

[Sithira @Sithira269]

@teej_dv @yacineMTB Wait what you got a terminal on remarkable ? Awesome was thinking of this to reduce my screen time

[Sithira @Sithira269]

does anyone in agentic engineering actually use an ice any more , I just use hunk to eye ball the diff

[Dimple Arora @DimpleAroro23s]

Name an IDE better than VS Code?