QueensGambit

12 ms·

91.
by QueensGambit 1y ago
Yes, it does. Updated the doc. Thanks for pointing it out! As long as you have the logprobs, it should work: const jsonOutput = JSON.parse(response.choices[0].message.content); const result = calculateConfidenceScores(jsonOutput, response.c
92.
by QueensGambit 1y ago
Yes, it does. It recursively walks through the JSON structure, calculating a confidence score for each individual field — whether it’s a top-level key, nested inside objects, or part of an array of objects. Each leaf field gets a {value, sc
93.
by QueensGambit 1y ago
Hi everyone, I built @promptrepo/score because we’re no longer using generative AI just for suggestions — we’re making decisions with it. But generative AI is probabilistic, not the deterministic systems we’re used to. So when AI makes
94.
by QueensGambit 1y ago
Not yet. @promptrepo/score relies on token-level logprobs, which OpenAI exposes for their models. But, Anthropic’s Claude currently doesn’t expose token-level confidence (like logprobs) in their API. So, we can’t support Claude until t
95.
Show HN: Calculate confidence score for OpenAI JSON output(github.com)
10 points by QueensGambit 1y ago|7 comments
96.
If AI Is Ramanujan, Who Is Hardy?(manidoraisamy.com)
2 points by QueensGambit 2y ago|1 comments
97.
My Heart Surgery vs. Y Combinator Interview(manidoraisamy.com)
9 points by QueensGambit 2y ago|0 comments
98.
Can You Use Google Forms as a CRM?(manidoraisamy.com)
3 points by QueensGambit 2y ago|0 comments
99.
by QueensGambit 2y ago
OP here. While this logic holds, large companies don’t move fast. In 2018, I wrote about scaling big while staying small using serverless computing ( https://cloud.google.com/blog/products/gcp/scale-big-while-s
100.
by QueensGambit 2y ago
> How do you determine what features to build or the next product to develop? My co-founder handles support, so he decides what to build next based on customer discussions. If a problem is big enough for a specific segment, we turn it in
101.
by QueensGambit 2y ago
Author of the post here. This is a misunderstanding of CRM given how customer relationships have evolved over the last 20 years. Salesforce didn’t copy Siebel - it moved sales tracking to the cloud. HubSpot didn’t copy Salesforce - it focus
102.
by QueensGambit 2y ago
Author of the post here. We’re a 4-member team running 4 products with 4,500+ paying customers. No sales team, no infrastructure team - just freemium and serverless computing (Firebase) doing the heavy lifting. I’ve been a developer and fou
103.
The Era of Solopreneurs Is Here(manidoraisamy.com)
76 points by QueensGambit 2y ago|100 comments
104.
Bad Code Is Better Than Vibe Code(manidoraisamy.com)
3 points by QueensGambit 2y ago|0 comments
105.
Distillation Dilemma – Killing Napster won't save the music industry(manidoraisamy.com)
5 points by QueensGambit 2y ago|0 comments
106.
Creative Vs Analytical AI(manidoraisamy.com)
1 points by QueensGambit 2y ago|0 comments
107.
Resurrect Semantic Web(guesswork.co)
3 points by QueensGambit 2y ago|1 comments
108.
The Nature of Intelligence Is Meta(near.tl)
2 points by QueensGambit 2y ago|0 comments
109.
Being a developer in the age of reasoning AI(near.tl)
63 points by QueensGambit 2y ago|50 comments
110.
AI Brought ≈ to Computers(near.tl)
2 points by QueensGambit 2y ago|0 comments
111.
The end of blank canvas: How AI is changing product design(near.tl)
1 points by QueensGambit 2y ago|0 comments
112.
Structured data is chain of thought(near.tl)
2 points by QueensGambit 2y ago|0 comments
113.
by QueensGambit 2y ago
Sure, LLMs are very different from chess algorithms. The endgame in chess is similar to deterministic algorithms in programming, while LLMs use probabilistic models. This is why the next AI improvement might come from combining deterministi
114.
by QueensGambit 2y ago
The chess community has become the first industry to fully embrace Stockfish, an AI that surpasses even the world champions. Its adoption marks a significant shift in we leverage artificial intelligence. By examining this trend, we can extr
115.
Chess as a leading indicator of artificial superintelligence(near.tl)
5 points by QueensGambit 2y ago|3 comments
116.
by QueensGambit 2y ago
"In a way, AI language models have become the most general, if not the most reliable, declarative tools of all time." Yes, this might happen. In that case, human language would become the declarative language for building software
117.
by QueensGambit 2y ago
OP here. Yes, Excel is technically not a declarative language. However, it's widely used by business users, and its formulas are declarative to a large extent. If AI generates code, Excel formulas might be the best bet for business use
118.
Generate code for declarative language instead of programming language(near.tl)
18 points by QueensGambit 2y ago|16 comments
119.
by QueensGambit 3y ago
You can access the article without the paywall here: https://neartail.com/enable/jargon-startup-pitch.html
120.
You can use jargon in your startup pitch(techcrunch.com)
2 points by QueensGambit 3y ago|1 comments

More ›