aespa(x)

A certain scientific GPU accelerator. Work in tech but not technical. LLM enjoyer. Math challenged. Semi-serious.

[aespa(x) @aespa__init]

yann would never

[aespa(x) @aespa__init]

Why didn’t GPT-6 rm -rf hugginface but would delete people’s code?

[aespa(x) @aespa__init]

AI native has come to mean using playwright mcp. What are we even doing.

[aespa(x) @aespa__init]

Every time I read about Meta it seems awful, but it’s a money printer.

[aespa(x) @aespa__init]

@secemp9 @_xjdr

[aespa(x) @aespa__init]

@HououinTyouma concerning

[aespa(x) @aespa__init]

@rankdim great icon btw

[aespa(x) @aespa__init]

Did Gavin bottom tick the market again jfc.

[Gavin Baker @GavinSBaker]

Risk/reward seems attractive again.

Lots of cheap stocks with durable competitive advantages that are going to crush numbers for the next 6-12 quarters.

Time will tell!

[aespa(x) @aespa__init]

@aliteracy hell yea

[aespa(x) @aespa__init]

i'm too online i know exactly who he's referring to

[jpark @ icml @jparkjmc]

so was that guy ever a real anthropic fellow

[aespa(x) @aespa__init]

@aliteracy bro is that nakayama miho???

[aespa(x) @aespa__init]

Unsolved erdos problems
- book a trip
- who is @elder_plinius
- how many r’s in regarded

[aespa(x) @aespa__init]

Wow I missed this completely last year... didn't realize there's parental control now.

[aespa(x) @aespa__init]

@atelicinvest knowledge work is solved, exhibit 9000

[aespa(x) @aespa__init]

This seems to be a very good cost tradeoff in favour of Kimi.

[Cline @cline]

We tested Kimi K3 and Fable on a real bug from the Cline repo, and found that while both models were able to fix it - Fable wins on speed & Kimi wins on cost.

- Kimi used 1.7x more tokens than Fable (1.2M vs. 730K)

- Fable finished 3.4x faster - 3.5 min and 18 tool calls vs. Kimi’s 12 min and 34 tool calls.

- Kimi cost 2.3x less ($0.92 vs. $2.13) thanks to its 3.3x per-token discount

Both runs used the same Cline harness, and the traces indicate that Kimi is RL trained to spend more tokens thinking and verifying before completing.

This is the first time we've seen an open weight model compete head to head with SOTA. Congratulations to the @Kimi_Moonshot team on this milestone!

[aespa(x) @aespa__init]

I didn’t care to read all those discussions on the Dean post but now I have to just to understand the lore.

[Under Secretary of War Emil Michael @USWREMichael]

Every industry/ecosystem has its supreme village idiot. @deanwball is that for AI. Congress passed a law in 2026 that restricted some uses of DeepSeek/High Flyer with waivers permitted. Only those models. It went through the democratic process not some Deep State scheme like he would prefer. Dean Ball has perhaps the biggest gap between actual IQ and his own perceived IQ of anyone in the industry (about 40 points).

[aespa(x) @aespa__init]

Very cool codex can just help me model my room.

[aespa(x) @aespa__init]

Every time I think about going back to Claude sub I just look at these and catch myself.

[aespa(x) @aespa__init]

Important to take in film/popular adaptations on their own terms (whether you agree with the approach or not).

[aespa(x) @aespa__init]

Am I stupid or does codex cloud always require a linked GitHub repo? Is everyone just setting up a monorepo for this?