Liam Dawson
@ldaws.com
Jack of all trades, master of pun. Board game fanatic. Software developer. He/him. Formerly @liamdaws on Twitter. #neurodiversesquad #adhd
Reposted by Liam Dawson
everyone forgets the full phrase
“fuck around, find out, write it down”
it’s only the scientific method if you write it down, otherwise it’s just sparkling curiosity
“fuck around, find out, write it down”
it’s only the scientific method if you write it down, otherwise it’s just sparkling curiosity
November 29, 2025 at 1:11 AM
everyone forgets the full phrase
“fuck around, find out, write it down”
it’s only the scientific method if you write it down, otherwise it’s just sparkling curiosity
“fuck around, find out, write it down”
it’s only the scientific method if you write it down, otherwise it’s just sparkling curiosity
Reposted by Liam Dawson
This is wild (via ajf) labs.watchtowr.com/stop-putting...
Stop Putting Your Passwords Into Random Websites (Yes, Seriously, You Are The Problem)
Welcome to watchTowr vs the Internet, part 68.
That feeling you’re experiencing? Dread. You should be used to it by now.
As is fast becoming an unofficial and, apparently, frowned upon tradition - w...
labs.watchtowr.com
November 26, 2025 at 9:56 PM
This is wild (via ajf) labs.watchtowr.com/stop-putting...
Reposted by Liam Dawson
jokes aside, can you imagine how you would handle it if you asked a human assistant "did you make all this up instead of actually doing any of the work", and that person responded to you with "good catch"
October 11, 2025 at 7:48 PM
jokes aside, can you imagine how you would handle it if you asked a human assistant "did you make all this up instead of actually doing any of the work", and that person responded to you with "good catch"
Reposted by Liam Dawson
"Glyph proposes that many LLM-assisted programmers who speak highly of the reliability and value of AI tools are falling prey to two cognitive biases: availability heuristic; and salience heuristic." /ht @benjaminbirnbaum
Pluralistic: LLMs are slot-machines (16 Aug 2025) – Pluralistic: Daily links from Cory Doctorow
pluralistic.net
August 16, 2025 at 11:08 PM
"Glyph proposes that many LLM-assisted programmers who speak highly of the reliability and value of AI tools are falling prey to two cognitive biases: availability heuristic; and salience heuristic." /ht @benjaminbirnbaum
Reposted by Liam Dawson
Funny to look back on the blog posts from AI CEO’s
(This is from 6 months ago)
www.businessinsider.com/anthropic-ce...
(This is from 6 months ago)
www.businessinsider.com/anthropic-ce...
Anthropic's CEO says that in 3 to 6 months, AI will be writing 90% of the code software developers were in charge of
"And then in 12 months, we may be in a world where AI is writing essentially all of the code," Anthropic CEO Dario Amodei said.
www.businessinsider.com
August 16, 2025 at 10:55 PM
Funny to look back on the blog posts from AI CEO’s
(This is from 6 months ago)
www.businessinsider.com/anthropic-ce...
(This is from 6 months ago)
www.businessinsider.com/anthropic-ce...
Reposted by Liam Dawson
Sam Altman has got to the point where he doesn't even bother to describe stuff anymore. "Uhh yeah sometime you'll have a thingy that will do stuff. And wow will it be good!"
August 6, 2025 at 2:57 AM
Sam Altman has got to the point where he doesn't even bother to describe stuff anymore. "Uhh yeah sometime you'll have a thingy that will do stuff. And wow will it be good!"
Chef's kiss (so to speak)
August 7, 2025 at 7:10 AM
Chef's kiss (so to speak)
Reposted by Liam Dawson
From the "rational arguments" away to the emotional ones, Alex put this beautifully succinct:
July 22, 2025 at 10:09 PM
From the "rational arguments" away to the emotional ones, Alex put this beautifully succinct:
Reposted by Liam Dawson
you can always do the “5 whys” in your incident review but be careful about going past that because after like 8 whys it’s just “capitalism”
May 17, 2025 at 9:49 PM
you can always do the “5 whys” in your incident review but be careful about going past that because after like 8 whys it’s just “capitalism”
Reposted by Liam Dawson
If your docs are shit (outdated, unclear, poorly organized), an LLM won't understand them any more than you do.
If your docs are *good* (written by knowledgeable people, with care to get it just right, maintained), an LLM gives you a subpar recap.
Invest in docs. Demand better docs. Read docs.
If your docs are *good* (written by knowledgeable people, with care to get it just right, maintained), an LLM gives you a subpar recap.
Invest in docs. Demand better docs. Read docs.
LLMs are good for coding because your documentation is shit
shkspr.mobi/blog/2024/07...
shkspr.mobi/blog/2024/07...
LLMs are good for coding because your documentation is shit
That's it. That's the post. Fine! I'll expand a little more. Large Language Models are a type of Artificial Intelligence. They can read text, parse it, process it using the known rules of English, a...
shkspr.mobi
July 12, 2025 at 9:00 PM
If your docs are shit (outdated, unclear, poorly organized), an LLM won't understand them any more than you do.
If your docs are *good* (written by knowledgeable people, with care to get it just right, maintained), an LLM gives you a subpar recap.
Invest in docs. Demand better docs. Read docs.
If your docs are *good* (written by knowledgeable people, with care to get it just right, maintained), an LLM gives you a subpar recap.
Invest in docs. Demand better docs. Read docs.
Reposted by Liam Dawson
Worth a watch:
Head of Signal, Meredith Whittaker, on so-called "agentic AI" and the difference between how it's described in the marketing and what access and control it would actually require to work as advertised.
Head of Signal, Meredith Whittaker, on so-called "agentic AI" and the difference between how it's described in the marketing and what access and control it would actually require to work as advertised.
June 26, 2025 at 4:28 PM
Worth a watch:
Head of Signal, Meredith Whittaker, on so-called "agentic AI" and the difference between how it's described in the marketing and what access and control it would actually require to work as advertised.
Head of Signal, Meredith Whittaker, on so-called "agentic AI" and the difference between how it's described in the marketing and what access and control it would actually require to work as advertised.
Reposted by Liam Dawson
The honest reason I don't have microtransactions/season pass/ads/100 DLCs/etc in Balatro isn't just about the ethics of those practices but because when I play other games that have those things it makes me want to put my computer in the dishwasher and set it to pots & pans
June 3, 2025 at 10:42 PM
The honest reason I don't have microtransactions/season pass/ads/100 DLCs/etc in Balatro isn't just about the ethics of those practices but because when I play other games that have those things it makes me want to put my computer in the dishwasher and set it to pots & pans
Reposted by Liam Dawson
Reposted by Liam Dawson
Reposted by Liam Dawson
Reposted by Liam Dawson
"Age verification" laws are actually "upload your ID or get your face scanned to access every website, ending anonymity and associating your identity with everything you do online" laws and if more people understood that they would not be down for this authoritarian nonsense
April 2, 2025 at 12:57 PM
"Age verification" laws are actually "upload your ID or get your face scanned to access every website, ending anonymity and associating your identity with everything you do online" laws and if more people understood that they would not be down for this authoritarian nonsense
Reposted by Liam Dawson
Addendum:
1. Vercel sat on this information for 22 days, not 5.
2. They have since rewritten the change log announcement I linked:
bsky.app/profile/edua...
1. Vercel sat on this information for 22 days, not 5.
2. They have since rewritten the change log announcement I linked:
bsky.app/profile/edua...
Yes, @vercel.com completely changed the changelog I linked to in my thread leaving no traces to the firewall marketing spin. And yes, I knew it would happen so I took a screenshot.
Here's the version of reality you'd be presented had they not been called out.
Here's the version of reality you'd be presented had they not been called out.
March 23, 2025 at 1:19 PM
Addendum:
1. Vercel sat on this information for 22 days, not 5.
2. They have since rewritten the change log announcement I linked:
bsky.app/profile/edua...
1. Vercel sat on this information for 22 days, not 5.
2. They have since rewritten the change log announcement I linked:
bsky.app/profile/edua...
Reposted by Liam Dawson
Columbia Journalism Review tested eight generative AI search tools and found their answers were wrong 60% of the time, and the paid ones actually fared worse than the free ones.
Meanwhile, millions of people trust the way they present total bullshit with confident language.
Meanwhile, millions of people trust the way they present total bullshit with confident language.
AI search engines cite incorrect sources at an alarming 60% rate, study says
CJR study shows AI search services misinform users and ignore publisher exclusion requests.
arstechnica.com
March 18, 2025 at 3:24 PM
Columbia Journalism Review tested eight generative AI search tools and found their answers were wrong 60% of the time, and the paid ones actually fared worse than the free ones.
Meanwhile, millions of people trust the way they present total bullshit with confident language.
Meanwhile, millions of people trust the way they present total bullshit with confident language.
1x.engineer - so good!
1x Engineer
The official website of 1x Engineers around the world.
1x.engineer
March 15, 2025 at 10:11 PM
1x.engineer - so good!
Reposted by Liam Dawson
its amazing how chatgpt knows everything about subjects I know nothing about, but is wrong like 40% of the time in things im an expert on. not going to think about this any further
March 8, 2025 at 12:13 AM
its amazing how chatgpt knows everything about subjects I know nothing about, but is wrong like 40% of the time in things im an expert on. not going to think about this any further
Reposted by Liam Dawson
A particular failure mode I have when trying to explain why something is a bad idea
If you list all 20 reasons, people will pick the 3 most-possibly-solvable and least-concerning ones, leaving the other 17 unaddressed
If you list the top 3, people will say "Well you can just bite those bullets"
If you list all 20 reasons, people will pick the 3 most-possibly-solvable and least-concerning ones, leaving the other 17 unaddressed
If you list the top 3, people will say "Well you can just bite those bullets"
March 3, 2025 at 9:40 PM
A particular failure mode I have when trying to explain why something is a bad idea
If you list all 20 reasons, people will pick the 3 most-possibly-solvable and least-concerning ones, leaving the other 17 unaddressed
If you list the top 3, people will say "Well you can just bite those bullets"
If you list all 20 reasons, people will pick the 3 most-possibly-solvable and least-concerning ones, leaving the other 17 unaddressed
If you list the top 3, people will say "Well you can just bite those bullets"
Reposted by Liam Dawson
Dr Hallowel advocating for the power of PLAY! Joy and playfulness are key to coaching ADHD
February 24, 2025 at 11:00 AM
Dr Hallowel advocating for the power of PLAY! Joy and playfulness are key to coaching ADHD
Reposted by Liam Dawson
They need him to be a genius because they cannot handle what it means for them to be tricked by a fool.
February 22, 2025 at 7:56 PM
They need him to be a genius because they cannot handle what it means for them to be tricked by a fool.
Reposted by Liam Dawson
Live versions of the "'What a week, huh?' meme.
- 'What a day': tintin.dlazaro.ca/day
- 'What a week': tintin.dlazaro.ca/week
- 'What a month': tintin.dlazaro.ca/month
- 'What a year': tintin.dlazaro.ca/year
#memes
- 'What a day': tintin.dlazaro.ca/day
- 'What a week': tintin.dlazaro.ca/week
- 'What a month': tintin.dlazaro.ca/month
- 'What a year': tintin.dlazaro.ca/year
#memes
February 18, 2025 at 4:07 PM
Live versions of the "'What a week, huh?' meme.
- 'What a day': tintin.dlazaro.ca/day
- 'What a week': tintin.dlazaro.ca/week
- 'What a month': tintin.dlazaro.ca/month
- 'What a year': tintin.dlazaro.ca/year
#memes
- 'What a day': tintin.dlazaro.ca/day
- 'What a week': tintin.dlazaro.ca/week
- 'What a month': tintin.dlazaro.ca/month
- 'What a year': tintin.dlazaro.ca/year
#memes