.

bitterer lesson pilled

[. @slopmath]

llm plateau

[Diogo Almeida @CompleteSkeptic]

almost like saying that there is a bitterer lesson than compute is all you need 🤔

https://t.co/bDZIJRpn39

[Eugenia Kuyda @ekuyda]

so i can book a flight and a hotel in 2 min myself OR listen to dots go through every flight option and send me screenshots of google maps on an excruciating 10 min call

[Adam Rackis @AdamRackis]

"Damn I wish instead of looking at a simple list of flight options I could make a phone call to an ai bot and verbosely describe my requirements and have an extended back and forth"

- Me, never

[. @slopmath]

@JTBeers @burkov you need a big model for huge context

[a16z @a16z]

98% of US households aren't paying for AI yet

More charts in State of Markets II: https://t.co/MTaxKUxa2w

[Diogo Almeida @CompleteSkeptic]

me to our GTM guy: do enterprises want benchmarks

him: no, the fast ones have already benchmarked their internal use cases. the slow ones are copying the fast ones.

he's onboarded a quarter of the fortune 500 already 🥵

[Ara Kharazian @arakharazian]

NEW: AI Spend FALLS in the latest Ramp AI Index.

Price cuts at the frontier, in addition to cheaper and more efficient models at standard and lite levels, are pushing down the cost of using AI.

Open source models remain <5% of business spend. This decline in spend is driven almost exclusively by competition between OpenAI + Anthropic.

[Unemployed Capital Allocator @atelicinvest]

Man, I really gotta get some work done.

But wait - should I use:

- ChatGPT web
- ChatGPT Work web - which is cloud environment for codex
- ChatGPT desktop?
- ChatGPT desktop work - which is local codex
- Codex - which is local codex
- Spaces - which is ... a mix?
- Projects?
- Launch a dot, which is basically cloud environment, but can do some extra stuff, but not really?
- Codex Cloud?
- Legacy Codex Cloud?

Man I wish I was making this up.
What is even going on anymore.

[Benjamin De Kraker @BenjaminDEKR]

This is slower and more friction than just normally booking the flight and hotel.

[. @slopmath]

has any company ever started shipping good products after hiring a bunch of ex meta PMs

[. @slopmath]

"build something people want" and we get a grating 10 minute phone call with an ai. I'll take clicking a few buttons thanks tho

[camsoft2000 @camsoft2000]

The phone call thing is a bit weird. It’s not a real phone call and doesn’t appear to use the phone call API on iOS which apps like WhatsApp and etc use. It’s just a simulated phone call using ChatGPT’s live voice mode feature.

[. @slopmath]

@tekbog I'm token rich and feel dumber than ever

[ben hylak @benhylak]

"haha OpenAI just TOTALLY killed Notion"

*sends tweet*
*opens Notion*

[. @slopmath]

@xlr8harder professional writing is still in the space of art. it comes naturally to most humans so we trivialize it, but defining it as a goal is something we've made no progress on (and far less progress than would be expected given where we're at on other seemingly less trivial things)

[. @slopmath]

@xlr8harder rlvr llms are really good at goalseeking in spaces where they have all of the inputs and the goal is explicitly defined as an optimization problem. nothing we know about the infinite+smooth space of art suggests it can be modeled as an optimization problem. goal can't be defined

[. @slopmath]

saying ai solved coding is like saying ghost kitchens solved food

[Valentin Ignatev @valigo]

Hey Claude - copy this one of a kind open source project. I'll go post how I made it in an hour on twitter to get likes! Copy it in a different language so it's a bit less blatant. Make no mistakes

[Rishabh Agarwal @agarwl_]

LLM crazy innovations be like:
- Please think step by step
- Please manage your context*

(*cries in KV cache)

[Y Disassembler @loomdoop]

Each successive tech boom just gets worse and worse.

[Hari @HarivanshRathi]

on ai psychosis:

as i watch engineers fall deeper into the belief that an amalgamation of mathematical probabilities somehow understands their codebase better than they do, i find myself thinking back to a time when software wasn’t built for hypergrowth, but simply to do x without inventing y.

ai seems almost fundamentally opposed to this philosophy. ask it to do x and it will eagerly invent y and z before it has even tried to understand x.

the danger isn’t that ai writes bad code- it’s that it makes writing unnecessary code 100% free

and the human condition is such that some of us will always prefer the fast, steep gains of ai, even when it does a bajillion unrelated things to accomplish something that could have been done without changing anything else.

so, somewhat paradoxically, the quality of software may keep declining for as long as ai keeps getting better. the cheaper complexity becomes to create, the less incentive there is to understand or avoid it.

try to preserve this craft created by our ancestors
write a LOC by hand today