Prashant Deva @pdeva

Geek. I like video games and math. Chief guy @ https://t.co/ZjunoadPpL San Francisco Joined April 2007

Tweets

9K
Followers

1K
Following

8K
Likes

85K

Prashant Deva @pdeva

2 days ago

Anthropic essentially had zero competition till recently. It will be interesting to see what they do now against real competition. Sonnet 5 for example has no place given its price/perf given the new competition.

0 0 0 87 0

View Details

Prashant Deva @pdeva

3 days ago

Fable limits are pointless frankly. You run into some random safeguard that will downgrade you to Opus long before you run into the usage limit..

0 0 0 54 0

View Details

Prashant Deva @pdeva

4 days ago

@theo keep 7 subs and now you have unlimited fable for each day of the week

0 0 1 440 0

View Details

Prashant Deva @pdeva

5 days ago

ran the exact same prompt on Opus 5, Fable 5, GPT Sol. Fable 5 == GPT Sol Opus 5 answers literally the opposite with terrible writing and poor explanation for its decisions. cant share the prompts/answers, but the above matches my experience in general with the models too.

0 0 0 65 0

View Details

Prashant Deva @pdeva

5 days ago

using kimi k3 to fix the slop Opus 5 created while i wait for Fable limits to reset..

0 0 1 80 0

View Details

Prashant Deva @pdeva

a week ago

JavaScript devs: "Rust is way too complicated." Also JavaScript devs: *rebuilds Rust’s type safety, error handling, concurrency model, and resource management in Effect.ts with 10x more complexity* : "All typescript code should use Effect.ts"

1 0 3 150 0

View Details

Prashant Deva @pdeva

2 weeks ago

all the bug tracking products are trying to become agent harness. but i just want a better bug database…

1 0 0 126 0

View Details

Prashant Deva @pdeva

2 weeks ago

Just yesterday, I was thinking that, wow, GitHub has really improved its uptime of late...

0 0 1 185 1

View Details

Prashant Deva @pdeva

3 weeks ago

@thsottiaux remove 5h limit permanently

0 0 0 9 0

View Details

Prashant Deva @pdeva

3 weeks ago

@levelsio if while using Fable it deems any task is “unsafe” , it will switch you to opus. this can happen a full hour into the sub agents working on your original prompt

0 0 1 199 0

View Details

Prashant Deva @pdeva

3 weeks ago

for pure chat, Gemini 3.5 flash remains the best model.

0 0 0 85 0

View Details

Prashant Deva @pdeva

a month ago

@AnthropicAI "coding and debugging will fall back to Opus 4.8" what even is the point of this model then? what are we supposed to use it for? generating bedtime stories?

0 0 7 565 0

View Details

Prashant Deva @pdeva

a month ago

@Jonathan_Blow any plans for a mac version? would love to play but no pc…

1 0 2 268 0

View Details

Prashant Deva @pdeva

2 months ago

tried the new screenx theatre at @AMCTheatres Metreon but the side panels never came on..... this is why going to movies sucks nowdays

0 0 0 69 0

View Details

Prashant Deva @pdeva

2 months ago

we need a $500 Claude Max plan for increased usage on Fable..

0 0 1 82 0

View Details

Prashant Deva @pdeva

2 months ago

@ericzakariasson I claimed mine, but it got associated with my team account... How do I have a single profile for both my personal and team account? GitHub allows one to do so easily by having multiple emails per github account.

0 0 0 82 0

View Details

Prashant Deva @pdeva

2 months ago

hot take: without @SBF_FTX leading Anthropic's Series B, the world would be a different place. No Anthropic means: - Openai has monopoly on frontier AI - token prices are 10x higher - coding AI revolution takes lot longer For a glimpse of that world, see O1 token pricing. Thats OpenAI operating without competition..