Eva
evaogbe.bsky.social
Eva
@evaogbe.bsky.social
they/them
software dev
foreva eva
Reposted by Eva
From another post (probably Twitter):
January 20, 2026 at 11:19 PM
Sans serif is doing me no favors today
My working theory is that Al is the new Mr Rogers. Famously religious nice guy that almost everyone feels positively toward.
I doubt there will ever be as good a song parodist as Weird Al.
January 21, 2026 at 5:20 PM
Reposted by Eva
The future of work is changing. And we have to evolve with it or fall behind. That's why we're proud to introduce: Serfdom dot AI. Our platform allows SERFs (solo entrepreneur revenue facilitators) to rent affordable housing that they can subsidize by picking up work with Baron, our AI assistant
January 20, 2026 at 11:56 PM
I’ve served enough time in the RxJS and Ramda trenches to know which way Effect is going to go
January 21, 2026 at 2:51 PM
Reposted by Eva
Your ChatGPT query uses 5928 tokens. To pay for these tokens, please purchase gems at the gem store. Gems can be bought with either mana crystals or premium Gold Coins
January 20, 2026 at 1:33 PM
Reposted by Eva
It's a crisis when you do it but it's impressive when a robot does it
January 20, 2026 at 1:54 PM
My main issue with Svelte is that if the Svelte team hasn't thought of a good way of doing something, then it's prevented entirely. The DX for the things they've thought through is pure bliss. The convoluted workarounds needed for the things they haven't anticipated are incredibly painful.
January 20, 2026 at 5:05 PM
Reposted by Eva
"you don't know how many people there are, but you can corral them all into the viewfinder of your camera at the same time" is one of those imaginary numbers invented by mathematicians to fuck with you
at last we will be free from the backbreaking drudgery of ordering pizza from a menu and even counting how many people the pizza is for www.theverge.com/tech/863365/...
January 17, 2026 at 11:16 PM
Reposted by Eva
for me, this may be the funniest product offering in the AI bubble so far
January 17, 2026 at 10:00 PM
Reposted by Eva
The myth persists that shipping reliable software is a "high bar".
January 18, 2026 at 11:46 AM
Reposted by Eva
It's funny that one of the things people learn as very small children is "you get better at doing the thing by doing the thing."

Everyone learns this as a small kid, it is very obviously completely true.

And using genAI manages to instantly make people forget it, and start arguing about it.
January 18, 2026 at 11:46 AM
Reposted by Eva
Morning
January 17, 2026 at 8:31 AM
Reposted by Eva
Thinking most of software industry management and influencers are suffering from some form of chatbot psychosis. Instead of "inventing" a new maths, they think they've invented a new form of coding, and because dev is so full of bullshit already people aren't noticing that it's outright unhinged.
January 18, 2026 at 5:20 PM
Reposted by Eva
Part of the struggle with the LLM discourse is that genuine (i.e. non-grifter) proponents talk exclusively about the possibilities of the transformer architecture in an ideal future with a rational business model.

Whereas opponents largely talk about the world we live in today, and its constraints.
What many of us have been saying for a while: AI*-related technologies are tools like any other technology, useful in some places, not in others. If it had been presented (and funded) as such, we'd be in a better place.

Instead, tech ideologues pushed AI* as the Second Coming, went all-in on it. /1
Insights from Ben Affleck on AI:

• AI can help write scenes but can't create full movies.

• Job loss fears are overblown because adoption of new tech is slow; it's hype for startup valuations.

• ChatGPT v5 is ~25% better but costs 4x.

• Users actually preferred v4's sycophancy for companionship.
January 18, 2026 at 5:38 PM
Reposted by Eva
Pretty much every non-hype defense of LLM products begins with "first you must already understand your work extremely well, have ironclad ethics, and also verify all of its outputs in their entirety" and these are simply not realistic conditions for a product to *require* before it can be useful.
January 18, 2026 at 5:51 PM
Reposted by Eva
January 19, 2026 at 9:55 PM
Reposted by Eva
I regret to inform7 you ma'am but your son is dead. Your son is found in the Grey Hallway. The description of your son is "a poor soul who died in these hallowed halls." The son is wearing a backpack. The backpack is openable and closed.
September 11, 2024 at 1:26 AM
Svelte's signals look like normal variables in an effort to simplify the mental model. But because they are not regular variables, you need to juggle when the signal needs to be wrapped in a getter. It would've been simpler to enforce calling signals like getters everywhere a la Solid.js
January 19, 2026 at 7:30 PM
I love this new era of fullstack type-safe RPCs. React, Svelte, Solid, and tRPC are all killing it. Finally we’ve stopped lying to ourselves that JSON APIs could ever or should ever be RESTful. Stringy types are dead. Long live functions!
January 17, 2026 at 10:50 PM
Reposted by Eva
You call it "over-engineering" when devs make software more complicated than it needs to be.

But simpler solutions often require *more* thought. Complexity's easy. You just keep typing.

That's why I call over-complicating "under-engineering".
January 16, 2026 at 8:50 AM
Reposted by Eva
Data is correct, causal agency is (IMO) incorrect.

The vast majority of ppl who use AI to make images and video have never touched and would never touch Photoshop/AfterEffects/Illustrator/etc. The issue is Adobe trying to squeeze its users with subscriptions, and other customer-hostile behavior.
From Dare Obasanjo on Mastodon. Adobe is hitting the floor. #AI
January 17, 2026 at 11:57 AM
Reposted by Eva
Nobody in a management position or the c-suite is willing to accept no amount of training data will actually FIX the issue with LLMs, because LLMs aren't minds. They don't learn or distinguish good info from bad. They return answers to queries that SOUND correct, given the training data provided.
January 17, 2026 at 11:49 AM
Reposted by Eva
FAQ:

What about those people who say they “don’t want any” but you know they’re gonna have some? Do they need to be in the group photo I upload to the papa John’s app? I don’t want the algorithm to fully count them, but also they need to be counted.
at last we will be free from the backbreaking drudgery of ordering pizza from a menu and even counting how many people the pizza is for www.theverge.com/tech/863365/...
January 17, 2026 at 3:03 PM
Reposted by Eva
I was not aware of bags until today, but apparently there is also a TanStack coin. It goes without saying that TanStack has nothing to do with it, will not claim any money and does not endorse buying this!

If you want to support open source, that's great, but not like this!
January 17, 2026 at 10:07 PM
Reposted by Eva
5 Accessibility Checks to Run on Every Component: "Hidde de Vries explains how to test components for accessibility, from keyboard support to screen readers and zoom." #a11y zeroheight.com/blog/5-acces...
5 accessibility checks to run on every component - zeroheight
Hidde de Vries explains how to test components for accessibility, from keyboard support to screen readers and zoom.
zeroheight.com
January 16, 2026 at 6:03 PM