I Think They Are Lying To You
354 segments
All right, people. It's time for a
passion yapping. I hope that you're
buckled in and sitting down because this
is going to be a bit intense. Now, I
want you to look right here. Do you see
this YouTube video? I'm not going to
play that YouTube video because
apparently someone has already been uh
DMCAD for showing any part of it. So,
I'm not going to have that happen. But
here's a quote from that video. Now it's
actually leveled up again to the next
abstraction where I don't prompt Claude
anymore. Now what what the heck are we
talking about? Well, this is Boris from
Anthropic, the man who started or
created Claude code talking about his
coding habits. Now 6 months ago, he
uninstalled his IDE and now he doesn't
even prompt Claude anymore. Okay,
prompting Claude is the old way we do
things. Now I set a goal. I let it run
in a loop. It burns just a cajillion
tokens until a win condition is met. And
bada bing, bada boom, we're done with
it. Okay, none of this hand coding. Ew,
gross. None of this prompting. Also ew
and also gross. Okay, we loop. And the
quote goes on. I have loops that are
running. They are the ones that are
prompting Claude and kind of figuring
out what to do. My job is to write
loops. And this is the kind of next
transition I think we're going to see in
the next few months, maybe through the
rest of the year. Now, this quote was
said and everybody, wow, deep insight.
And everybody kind of clapped about it.
And if you haven't been following
Boris's journey, he's also said phrases
a little while back, coding is mostly
solved. Now, he just says coding is
solved. And even more, he says coding is
the easy part. The hard parts like
infrastructure or hardware or uh
listening to users or getting feedback.
Those are the hard parts. But coding
easy peasy. It is solved. Lemon squeezy.
Not a big issue. Now I take a little bit
of issue with that. Okay. In fact, I
would like to say that I take well a lot
of bit of issue with that. And let me
kind of build up a case and I'm going to
explain to you why I think they're all
lying. I think they are all lying to
you. And I know that's going to kind of
ruffle some feathers. Well, first off,
you need to see this tweet right here
cuz it's very important as part of this
entire concept. This is Anthropic
saying, "Hey, we've greatly accelerated,
okay? We are faster. We are stronger.
We're better than we've ever been
before. In fact, in this last quarter,
Q2 2026, we are shipping 8x the amount
of code per employee than we were on the
average before 2025." In other words,
they are shipping two years worth of
code every single quarter for every
single employee. Hence the reason why
coding is the easy part. Coding is
solved, bro. We don't we don't write
code no more. Okay? We prompt the
prompters. All right? Big difference. We
write loops. We do not code ourselves.
Review code. It's all done
automatically. Hey, me from the future
here. So when I say they are lying, what
I mean is either A they are
intentionally lying to you or B they so
believe what they say that even in the
face of contrary evidence they go now
I'm correct. I no I am right it doesn't
Nope. The sky is in fact not blue. No
it's not will not accepted. Tell me I'm
wrong. I don't think I am. All right
back to past me. By the way I get even
more passionate. So you're welcome. I
have been using linear lately. And look
at this. Okay. I just clicked on a
ticket. It loaded. I go back. I go
forward. Look at how fast this is. I can
drag them between columns. Absolutely
creamy smooth. I know this is unusual.
This is not what you expect from
software, but this is some of the best
software I've been using, and I use it
to track all the things I am working on,
and you should, too. You should try out
Linear at linear.app. Now, to the main
thesis. I think they are lying, and I
think that they're hurting people. I
think they're genuinely hurting people.
First off, I get lots of messages about
people genuinely worried about the
future and hey, maybe there's no room
for me left. Hey, I think everything's
going to be destroyed. People's jobs are
becoming this really frustrating
experience where they just have people
yeeting code into production. There's
absolutely no controls because this is
what they've been told to do. Run the
loops, code the coders, just let it run.
Let it fly. And so, just even on a
personal level, people are feeling
really burnt out and freaked out. At
least that's the general gist that I'm
getting from a ton of people. This has
not been so good. And a lot of people's
well-being and how kind of their outlook
in life has actually gone very negative.
I think what anthropic and more
specifically Boris is doing is very
destructive. Now, I know I'm going to
get, "Hey, man, Boris, he's a he's a
nice guy." Okay? Like, in person, very,
very nice. And you know what? I'm sure
you're right. In person, humble and
nice. But here's the deal. humble and
nice people can do bad things. Now, let
me make the case to you why I think that
they are lying and I think it's very
very obvious. So, the first thing you
know to understand is that Claude code
was released for researchers in February
2025. Now, if you're unfamiliar, Claude
Code is a little terminal application in
which you can prompt Claude to go and do
some angentic work and produce code. All
right. Well, within two weeks of
releasing that GitHub issue 392, hey
bro, screen's flickering. A little bit
afterwards, April 11th, 2025 in
progress. Screen's flickering,
everybody. And this continues on. 1913,
terminal flickering, and there's like so
many other issues devoted to terminal
flickering. It's almost as if Anthropic
thought people needed a little bit more
epileptic training so that they can
watch their favorite animes. But
nonetheless, this is a very well-known
issue and it was reported almost
immediately upon research release.
Meaning that this bug most likely
existed not even in 2025, potentially in
2024 while it was all being developed
and they actively knew about it is my
personal guess. If you used the product
at all during that time period, it was
so evident. I remember having my own
experiences with it, being shocked
people even used the product at all.
Well, let's fast forward a little bit.
They finally respond to it publicly
December 17th, 2025.
And I quote, "We've rewritten Claude
Code's terminal rendering system to
reduce flickering by roughly 85%." Now,
first off, that's impressive. I have
never ever in my lifetime seen a bug fix
that only works 85% of the time. Like,
what kind of stochastic process are is
going on out there? Like, that is
insane. And then they go on to explain
how difficult it was. They're building a
game engine brothers to render some text
to a terminal. This was said out loud
couple days later, December 18th. We're
rolling back a few changes we shipped to
Claude Code this week to make sure
things are stable heading into the
holidays. In other words, there, hey,
coding is largely solved. Turns out it's
not largely solved. Now, you're probably
wondering, why am I harping so hard on
this screen flickering issue? Of course,
being known for like nine months at that
point, why am I harping so hard on it?
Well, here's the thing. Screen
flickering, that's a pure software
problem. There's no hardware involved in
this. Okay? There's no, "Oh, the
infrastructure wasn't quite quite right.
Oh, the capacity. We just didn't have
the capacity for non-flickering terminal
experience." No, it's the user feedback.
We just didn't have the feedback. We
didn't even know users didn't like
flickering. No, they knew. They were
well aware. They probably even
prioritized a flick flickering
considering they also did a whole
rewriting of the rendering system. But
ultimately the first swing, the first
rewrite failed failed epically failed
right next to Christmas and they had the
poll out. And so hey, is software
largely solved? I'm not buying it
currently. But here's the thing is if
the story ended there, maybe we could
move on, but it doesn't. Okay, March
25th, 2026. Now we are officially over a
year into first known reported issue on
GitHub. Over a year. Just let that sink
in. Over a year of the terminal
flickering. Text characters laid out in
a grid. Couldn't help but to flicker.
I
do. I feel crazy that this is a crazy
hard problem. No. This cannot I refuse
to believe that this is that hard of a
problem. Well, guess what? It was
finally asked, "Please fix the
uncontrollable scrolling/flickering
before the next 3,000 features." People
were sick of it. And on April 1st, not
as a joke, Boris releases no flicker
mode for Claude Code. This is that new
rendering that they talk about. It's
called Claude Code, no flicker. I
believe it takes advantage of alternate
screen as opposed to doing this direct
print down because the direct print down
does take a little bit more effort to
get right. Whereas alternate mode,
that's like how Vim works, right? It's
beautiful. I I love I love Vim. Okay.
Hey, we're talking about Vim. I could I
could I could sing some praises about
Justin and his running of Neo Vim. But
back to the thing. I want you once again
to think about this. This was only two
months and a little bit of change ago
and they still haven't fixed flickering.
It's been over a year and they have a
feature branch or a feature flagged
feature in which solves flickering via a
completely different path. Now, I want
you to think about that. If coding is
solved, do you need feature branches?
No.
No. No. No. No. You don't. Somebody's
lying to you. Because if coding was
easy, if coding was solved, no, you
don't need that. Who is telling the
truth right now? Because it does not
seem like Anthropic or Boris is leading
you to the correct path. Now, I don't
know if they're actually actively lying
to you, but in my head, I feel like they
are. To me, it seems apparent that I
think they are liars. Now, lastly, we
get all the way up to May 27th with
Claude Dev's official account even
saying, "Hey, guess what? First off, I
just want to let everybody know. New
terminal dropping, right? This is the
official Claude Dev account saying,
"Hey, we're doing it. We're dropping the
big deal." But even more so, if you
scroll a little bit down, you'll see
this fewer mysterious error messages.
They have error messages in a product
that they're just like, "I don't know.
I guess it's not working right now.
We're not really sure what the heck's
happening." Are you kidding me? Yo, bro,
why don't you launch a loop? Why don't
you launch a loop like 6 months ago?
Bro, write a loop and fix that. Don't
have a mysterious error. Let people
know. Oh, whoopsies. Turns out the
connection terminated. You would you
like to try again or would you like to
start over? Oh, hey. Whoopsies. It turns
out that you're out of credits. Sorry,
you can't continue. Like, there's only
so many problems here. Write out the
problems and the errors. Loop it. Loop
it, brother. Anyways, I say all this
because even as I'm confused about this
entire thing, Boris even responds to me
saying, "Coding is the easy part.
Everything else is not yet solved, but
is also becoming increasingly
automated." saying to me, "Coding is
solved.
It's the easy part." And yet during all
of this, it just feels like they have an
abundance of obvious failures. And if
you just look at their clawed status,
their claw status, it's like at 98 some
percent up and they're having just
errors elevated with this model, errors
with this model. Like they're not like,
"Oh, we ran out of capacity. Sorry,
overloaded with requests." No, it's
like, bro, we actually like there's
actual errors. Not only that, but for
the last year, you can see GitHub
issues, Reddit posts, hacker news posts
saying, okay, I was prompting and it
looks like I got somebody else's
feedback or I got somebody else's prompt
result cuz I'm getting like a legal
document, but um there's none of us are
talking about legal documents. What the
heck's going on, Claude? And there's
this whole thing where people are
accusing them of leaking other people's
session. And I'm not even saying they
are or they aren't or that this is some
crazy hallucination case, but they can't
and don't know how to solve that issue.
They don't even really know what's
happening with that one because it
continues to happen. And so it's like,
yes, is coding hard? Sure, coding can be
hard. That's okay to say. No, nobody's
going to be upset about that. Have we
accelerated the rate in which we can
produce code? Yes. Have we accelerated
the rate in which we could produce more
correct code? Do we have tools in which
can help people write code? Sure. Yes.
Beautiful, glorious, knock yourself out.
But this idea that everybody should be
writing loops, everybody should be
spending $10,000 a day on my company
salesman selling his items and just
telling everybody this is the way it
should be when they get their tokens
free and unlimited while at the exact
same time making demonstrabably false
claims. Drives me bonkers. And I just
had to talk about it. because they in my
opinion they're lying to your face. They
are peeing on your leg and saying it's
raining. I can hear it's raining right
now. That's why I said that. Anyways,
that's the video. That's what I wanted
to say. Hey, the name is the primogen
Ask follow-up questions or revisit key timestamps.
The video presents a critical perspective on the claims made by Anthropic and Boris, particularly the assertion that 'coding is solved' and that developers should rely on autonomous loops instead of manual coding. The speaker argues that these claims are misleading and potentially harmful, citing persistent technical issues with the Claude Code tool—such as terminal flickering and mysterious errors—as evidence that coding remains a complex and unsolved challenge. The speaker questions the integrity of these claims, suggesting that they are being used to push a narrative that prioritizes automated output over stable, high-quality software development.
Videos recently processed by our community